Granular composite materials

JP7580340B2Active Publication Date: 2024-11-11THE TRUSTEES OF PRINCETON UNIV +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
JP2021092969
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Priority Date
2014-06-20
Filing Date
2021-06-02
Publication Date
2024-11-11
Estimated Expiration
2034-08-18

AI Technical Summary

Technical Problem

Existing methods struggle to effectively predict and control the structure of granular composites, leading to inconsistent physical characteristics such as porosity and viscosity, which affects the performance of materials like concrete and solid propellants.

Method used

A method involving the compounding of distinct particle groups with specific volume probability density functions and mixing conditions to create a Dense Small-size-range Multimodal Granular (DSMG) structure, reducing porosity and viscosity through precise particle group combinations.

Benefits of technology

The DSMG structure achieves lower porosity and viscosity, enhancing properties like bulk modulus, thermal conductivity, and processing integrity in materials like metals, ceramics, and concrete.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 00000045_0000
    Figure 00000045_0000
  • Figure 00000045_0001
    Figure 00000045_0001
  • Figure 00000046_0000
    Figure 00000046_0000
Patent Text Reader

Abstract

To provide a system for reducing the porosity of a granular composite to improve physical properties of a composite; a layer of the granular composite; and a layer-forming method.SOLUTION: A system includes: a) a dispenser positioned over a target surface, the dispenser containing a powder-like or suspension-like granular composite having a porosity of 20% or less; and b) an energy source positioned to transfer energy to the composite when the composite is dispensed on the target surface.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art

Description

[Technology Field]

[0001] This invention was made possible with government assistance under grant numbers DMR-0820341 and DMS-1211087 granted by the U.S. National Science Foundation. The government has certain authority over this invention. [Background technology]

[0002] Granular materials, or granular composite materials, are aggregates of constituent particles, each having a defined geometric shape (size and form) such that it remains substantially fixed when placed adjacent to other components and pressed against one another, for example, by gravity. In granular composite materials, the constituent components may be suspended in a solvent or liquid, or their positions may be roughly or precisely fixed by a “paste” or “adhesive component.” Granular composite materials are ubiquitous in industry, laboratories, and nature. Common examples of naturally occurring granular composite materials include soil, sand, and gravel, while common examples of artificial granular composite materials include concrete, bird hunting pellets, sugar, baby powder, solid propellants, cermets, ceramics, inks, and colloids.

[0003] The physical properties of a granular composite material depend closely on the fine many-body structure formed through the physical interactions of its constituent particles, and on the physical properties of the materials that make up its components. These properties include, but are not limited to, porosity (the gaps not filled by the constituent particles), viscosity, mechanical strength, toughness, tensile strength, elastic modulus, bulk modulus, shear modulus, thermal conductivity, electrical conductivity, and coefficient of thermal expansion. For example, composite materials made of certain types of materials with high porosity are generally weaker in strength, more thermally conductive, and more electrically conductive than composite materials made of the same type of material with low porosity. Alternatively, composite materials made of very rough (high coefficient of friction) and aspherical components, when mixed irregularly, generally form a lower-density (higher porosity) structure compared to composite materials made of the same material components that are relatively smooth and aspherical.

[0004] Research on granular composite materials and their applications has generally focused on measuring both the physical properties of the composite material in question and the geometric size, shape, and other physical properties of its constituent particles. For example, in the concrete industry, when crushed stone and sand are mixed with water-containing cement ("paste") in specific proportions to form concrete, a "pass-through curve" is frequently used to roughly represent the size distribution of the constituent particles in the mixture. This "pass-through curve" is obtained by passing a dry mixture of sand and crushed stone (also called aggregate) through a series of finer and more continuous sieves, and then plotting the volume (or mass) fraction of the aggregate that passed through each sieve. For example, it is known that by changing the particle size distribution, for instance, by reducing the amount of smaller-sized aggregate (in this case, sand), the physical properties of water-containing concrete, such as its viscosity, and also the physical properties of dry and set concrete, such as the elastic modulus and durability of the concrete, can be altered. In this way, some researchers have sought to improve the properties of concrete by changing the mixing ratio of aggregates. F. de Larrard, Concrete optimization with regard to packing density and rheology, 3rd RILEM international symposium on rheology of cement suspensions such as fresh concrete, France (2009). JMShilstone, Jr., and JMShilstone, Sr., Performance based concrete mixtures and specifications for today, Concrete International, 80-83, February (2002).F. de Larrard, Concrete Mixture Proportioning, Routledge, New York (1999). J. M. Schilstone, Concrete Mixture Optimization, Concrete International, 33-40, June issue (1990).

[0005] However, the wide range of problems in designing granular composite materials based on the geometric shapes and characteristics of their constituent components has generally been difficult to handle due to its immense complexity. The characteristics of a composite material depend not only on the fine geometric shapes and physical characteristics of each and all of its constituent components, but also on the position, orientation, and arrangement of all particles in the composite material. For example, the structure of a composite material obtained by shaking its components in a sealed container and then pouring it into another container will have different porosity than the structure produced by the exact same components vibrated at a high frequency in the container. This difference can be extremely significant, for example, resulting in a porosity that is 50% lower than that of the vibrated preparation, and the essential differences between structures with different porosities will have a significant impact on the physical characteristics of the composite material.

[0006] In concrete, for example, it has been shown that the mechanical strength of concrete depends exponentially on the porosity of the aggregate mixture, with mixtures exhibiting lower porosity being exponentially stronger. However, viscosity also depends exponentially on the porosity, inversely proportional to flowability, with mixtures exhibiting lower porosity being less fluid (having higher viscosity). Concrete must be fluid to be poured over a certain area on site, and therefore, even if a higher porosity means lower strength, a higher porosity will be required in the aggregate mixture. Another example is granular armor, where a low porosity in the armor before molding means high viscosity, and even if the strength increases, it will make the final armor more difficult to process. Regarding solid accelerators, the thrust of a rocket generally depends on the square of the density of the propellant in its composite material (the density of the composite material is proportional to 1 minus the porosity).

[0007] Generally, the ability to effectively predict, design, and control the structure of granular composite materials is essential for broadly managing the physical properties of composite materials. Of particular importance is the ability to reduce the porosity of the composite material to improve its physical properties, and often while maintaining a low viscosity sufficient to preserve the capabilities required for processing. [Overview of the project] [Means for solving the problem]

[0008] The present invention relates to increasing the density of granular composite materials, and related methods and compositions. Applications in which these properties are beneficial include, but are not limited to, 1) additive manufacturing ("3D printing") involving metals, ceramics, cermets, polymers, plastics, or powders or gels that are dry or suspended in a solvent; 2) concrete materials; 3) solid propulsion materials; 4) cermet materials; 5) granular armor materials; 6) glass-metal and glass-plastic mixtures; and 7) ceramics containing (or manufactured using) granular composite materials.

[0009] In one embodiment, the present invention contemplates a compounding method for producing a material with a low porosity. In one embodiment, the present invention contemplates a method for manufacturing a particulate composite material composition comprising: a) providing at least first and second distinct groups of at least 100 particles, each group representing a volume probability density function P i (V) of the particles in that group and having one or more maxima, an average particle size V i avg and a passing curve, and b) mixing the particles from two or more of the foregoing groups under the condition that a combination of at least 50 particles from each group produces a combined particulate composite material exhibiting a combined passing curve representative of the volume probability density function P(V) of the combined particles, wherein the combined particulate composite material has the following characteristics: i) associated with different particle groups "1" and "2" having a V2 avg / V1 avg ratio of 10,000 or less at at least two maxima, having a maximum at a smaller volume (point) indicated by V1 and a maximum at a larger volume indicated by V2, ii) having at least one minimum between the two aforementioned maxima, the height of the passing curve of the minimum not being higher than 75% of the height of the passing curves of both maxima, and iii) having positive direction points V l and V r with at least one of the aforementioned minima between them and V r / V l = 10,000, such that the integral value of V*P(V) from V l to at least one of the two aforementioned maxima and the minimum conforming to criterion ii) between them is at least 2% of the integral value of V*P(V) from V l to V r , and such that the integral value from the same minimum to V r is at least 2% of the integral value from V l to V r . In one embodiment, V2 avg / V1 avg <= 2,000 and >= 25.

[0010] In one embodiment, there are two or more particle groups and the method includes dividing the particle groups into subsets prior to step b). In one embodiment, the mixing is carried out under conditions that inhibit phase separation. In one embodiment, the combined composite material exhibits a porosity of less than 25%, or less than 20%, or even less than 15%. In one embodiment, the mixing reduces the (relative) viscosity. In one embodiment, the combined composite material is immersed in a solvent, paste, gel, liquid, or suspension. In one embodiment, V2 avg / V1 avg The values ​​are (<=) 2,000 or less and (>=) 25 or more. In one embodiment, the method further includes optimizing a procedure for calculating a low porosity combination using two or more of the aforementioned particle groups, wherein the optimization is performed on a porosity function P corresponding to the mixing of particle groups. i ( i , j1、 φ i , j2 ···φ i , jN ) or partial subset porosity function P j1,j2, ···(φ i,j1、 φ i,j2...) is obtained. In one embodiment, the optimization procedure includes constraints on the physical properties of the combined composite material. In one embodiment, the aforementioned mixing results in a very integrated structure. In one embodiment, the aforementioned mixing results in a nearly very integrated structure. In one embodiment, the aforementioned mixing is carried out in a container having a diameter and height at least 100 times that of the largest particle of its largest group. In one embodiment, the aforementioned mixing includes adding particles from its second group to the particles of the first group. In one embodiment, the aforementioned mixing results in a final proportion of particles from groups having a smaller average particle volume of approximately 10-80%. In one embodiment, the aforementioned mixing includes adding particles from its first group to the particles of the second group. In one embodiment, the particle size distribution corresponding to the aforementioned first and second groups exhibits an arithmetic standard deviation of less than 20%. In one embodiment, the method further includes c) using the aforementioned granular composite material having a final porosity of less than 25% as a powder, for example, as a powder for laser sintering, or as a powder for laser melting, or as a powder for additive manufacturing of ceramics, or as a powder for powder metallurgy, or as a powder for injection molding, or as a powder for the production of granular armor, or as a powder for some other purpose. In yet another embodiment, the method further includes c) using the granular composite material in concrete manufacturing.

[0011] The present invention also considers compositions produced by the methods described herein. In one embodiment, the present invention provides a granular composite material composition comprising at least 100 particles, wherein the passing curve represents the probability density function P(V) of the particle size (volume), i) volume V i and V j At least two local maximums that appear, the smaller volume side is V i It is displayed as and its large volume side is V j It is displayed as follows: ii)V i <V i-j <V j V isi-j It has at least one local minimum that appears in V i-j The height of the passing curve for the local minimum that appears is not higher than 75% of the height of the passing curve for both local maximums, iii) V h-i <V i <V j <V j-k Point V h-i and V j-k There is a point V h-i This is the volume of the smallest particle in the composite material or V h and V i The greater of the two local minima between the local maximums and V h <V i Therefore, V h The maximum value at corresponds to particle group "h" that satisfies all criteria i), ii), iii), iv), and v), and here point V j-k This is the volume of the largest particle in the composite material or V j and V k It is smaller than either of the local minimums between the local maximums in V j <V k It is determined that V k The maximum value in corresponds to particle group "k" that satisfies all criteria i), ii), iii), iv), and v), and corresponds to particle groups "i" and "j" in iv), and V j avg / V i avg Average particle volume V ≤ 10,000 i avg and V j avg And here, group "i" is V h-i From V i-j It is defined as a group containing all particles having volumes within the range of V, and group "j" is V i-j From V j-k It is defined as a group containing all particles having a volume within the range of, and v) point V l and V r There is, V l <V i avg <Vj avg <V r and V r / V l = 10,000, and thus V h-i from V i and V j the minimum value V between i-j at least one of which meets criterion ii) and the integral value of V*P(V) up to the minimum value that meets criterion ii) is at least 2% of the integral value of V*P(V) from V l from V r to V i-j and thus the integral value from the same minimum value at V j-k to V l from V r to V is at least 2% of the integral value of V*P(V) from V j avg > V i avg and V j avg / V i avg<=2,000 and >=25. In one embodiment, for at least one pair of adjacent (by average volume) particle groups, the average number of closest large particles in the larger (by average volume) group of the two groups is 1 or greater. In one embodiment, the relative volume of the smaller (by average volume) group of at least one pair of adjacent (by average volume) particle groups is between 10% and 80% of the total volume of particles in that pair group. In one embodiment, a mixture of fixed amounts of particle groups occupies a larger volume space than any single particle group occupies by its own fixed amount. In one embodiment, the aforementioned composite material exhibits a porosity of less than 25%, more preferably less than 20%, and even more preferably less than 15%. In one embodiment, the aforementioned granular composite material is immersed in a solvent, paste, gel, liquid, or suspension. In one embodiment, spatial phase separation of particles into groups of similar size does not occur for all particle groups. In one embodiment, the pair correlation function of the granular composite material described above represents the probability increase of a linear arrangement of the central part of three contacting particles, where two particles have a volume at least 25 times that of the remaining particles, or two particles have a volume at least 25 times smaller than that of the remaining particles.

[0012] In one embodiment, the present invention considers systems, layers, and methods corresponding to additive manufacturing. In one embodiment, the present invention considers a system comprising a) a granular composite material having a porosity of 20% or less, an extruder (dispenser) positioned above the surface of a target object, and b) an energy source positioned to supply energy to the composite material when the composite material is extruded onto the aforementioned surface of the target object. In one embodiment, the aforementioned composite material is protected from oxidation by an inert gas. In one embodiment, the aforementioned composite material is a sinterable powder. In one embodiment, the aforementioned composite material is a fused powder. In one embodiment, the aforementioned composite material is a molten powder. The present invention is not intended to limit itself to powders having a specific melting temperature. In one embodiment, the aforementioned molten powder exhibits a melting temperature of 500°C to 5,000°C. The present invention is not intended to limit itself to the characteristics of particles used to produce the powder. In one embodiment, the aforementioned powder contains metal particles. In one embodiment, the aforementioned powder includes ceramic particles. In one embodiment, the aforementioned powder includes cermet particles. In one embodiment, the aforementioned powder includes a mixture of ceramic particles and metal particles. In one embodiment, the aforementioned powder includes carbide particles. In one embodiment, the aforementioned powder includes glass particles. The powder may be a mixture of two or more particle types. In one embodiment, the aforementioned powder includes a mixture of polymer particles and metal particles. In one embodiment, the aforementioned powder includes a mixture of polymer particles and ceramic particles. In one embodiment, the aforementioned powder includes a mixture of polymer particles and glass particles. In one embodiment, the aforementioned powder includes a mixture of metal particles and glass particles. In one embodiment, the aforementioned powder includes a mixture of carbide particles and polymer particles. In one embodiment, the aforementioned powder includes a mixture of carbide particles and metal particles. In one embodiment, the aforementioned powder includes a mixture of carbide particles, cermet particles, and metal particles. In one embodiment, the aforementioned powder comprises a mixture of carbide particles, cermet particles, and polymer particles.In one embodiment, the aforementioned powder comprises a mixture of ceramic particles, metal particles, and polymer particles. In one embodiment, the aforementioned powder comprises a mixture of metal particles, glass particles, and polymer particles. In one embodiment, the aforementioned powder comprises a mixture of carbide particles, metal particles, and polymer particles. In one embodiment, the aforementioned powder comprises titanium alloy particles and has a porosity of approximately 10%. In one embodiment, the aforementioned powder comprises first and second groups of particles, wherein the first group of particles has an average particle volume at least 25 times larger than the average particle volume of the second group of particles. In one embodiment, the aforementioned powder comprises first and second groups of particles, wherein the first group of particles has an average particle volume between 25 and 2,000 times larger than the average particle volume of the second group of particles. In one embodiment, the aforementioned powder comprises a mixture (by volume) of a first group containing particles of about 10 microns, a second group containing particles of about 2 microns, a third group containing particles of about 200 nanometers, and a fourth group containing particles of about 40 nanometers, each in the proportions of 62.8%, 16.2%, 16.7%, and 4.3%, and the aforementioned powder has a porosity of about 4.4%. In one embodiment, the aforementioned energy source is a laser. In one embodiment, the aforementioned granular composite material is in a solvent. In one embodiment, the aforementioned granular composite material is in a paste.

[0013] The present invention also considers layers. In one embodiment, the present invention considers a layer of powder in a granular composite material, the layer having a thickness of less than 1,000 microns, and the powder having a porosity of 20% or less. In one embodiment, the aforementioned powder is easily sinterable. In one embodiment, the aforementioned powder is fused. In one embodiment, the aforementioned powder is meltable. Again, the present invention is not intended to limit itself to the characteristics of the particles used to produce the powder. In one embodiment, the aforementioned powder includes metal particles. In one embodiment, the aforementioned powder includes ceramic particles. In one embodiment, the aforementioned powder includes cermet particles. In one embodiment, the aforementioned powder includes carbide particles. Again, the present invention considers mixtures of particles. In one embodiment, the aforementioned powder includes a mixture of ceramic particles and metal particles. In one embodiment, the aforementioned powder includes a mixture of ceramic particles, metal particles, and polymer particles. In one embodiment, the aforementioned powder comprises titanium alloy particles and has a porosity of about 10%. In one embodiment, the aforementioned powder comprises first and second groups of particles, the first group having an average particle volume at least 25 times larger than the average particle volume of the second group of particles. In one embodiment, the aforementioned powder comprises first and second groups of particles, the first group having an average particle volume between 25 and 2,000 times larger than the average particle volume of the second group of particles. In one embodiment, the aforementioned powder comprises a mixture of a first group containing particles of about 10 microns, a second group containing particles of about 2 microns, a third group containing particles of about 200 nanometers, and a fourth group containing particles of about 40 nanometers, each in a ratio of 62.8% to 16.2% to 16.7% to 4.3%, and the aforementioned powder has a porosity of about 4.4%. In one embodiment, the aforementioned layer is placed on a second layer of granular composite material powder, the second layer having a thickness of less than 1,000 microns, and the powder having a porosity of 20% or less. In one embodiment, both layers have a thickness of about 50 microns. In one embodiment, the aforementioned granular composite material powder is in a solvent.In one embodiment, the powder of the aforementioned granular composite material is present in a paste.

[0014] The present invention also considers a method for creating layers. In one embodiment, the present invention considers a method for producing layers comprising the steps of: a) providing a source of granular composite material having a porosity of 20% or less; b) depositing a first portion of the aforementioned powder onto the surface of an object; c) accumulating energy in the first portion of the powder under conditions such that the energy causes the first portion of the powder to sinter, fuse or melt, thereby creating a first layer; d) depositing a second portion of the powder on the aforementioned first layer; and e) accumulating energy in the aforementioned second portion of the powder under conditions such that the energy causes the second portion of the powder to sinter, fuse or melt, thereby creating a second layer that is placed on the aforementioned first layer. The present invention is not intended to limit itself to an energy source. In one embodiment, the energy is accumulated by a laser. In one embodiment, step c) includes focusing the laser with at least one lens. The present invention is not intended to limit itself to layers of a specific thickness. However, in one embodiment, the aforementioned first and second layers are 100 microns (or less than 50 microns) thick. The present invention is not intended to be limited by the characteristics of the particles used in the production of the powder. In one embodiment, the aforementioned powder includes metal particles. In one embodiment, the aforementioned powder includes ceramic particles. In one embodiment, the aforementioned powder includes cermet particles. In one embodiment, the aforementioned powder includes carbide particles. In one embodiment, the aforementioned powder includes a mixture of ceramic particles and metal particles. In one embodiment, the aforementioned powder includes a mixture of ceramic particles, metal particles and polymer particles. In one embodiment, the aforementioned powder includes titanium alloy particles and has a porosity of about 10%. The present invention is not intended to be limited by the number of particle groups. In one embodiment, the aforementioned powder includes first and second particle groups, wherein the particles of the first group have an average particle volume at least 25 times larger than the average particle volume of the particles of the second group. In one embodiment, the aforementioned powder includes first and second particle groups, wherein the particles of the first group have an average particle volume between 25 and 2,000 times larger than the average particle volume of the particles of the second group.In one embodiment, the aforementioned powder comprises a mixture of a first group containing particles of about 10 microns, a second group containing particles of about 2 microns, a third group containing particles of about 200 nanometers, and a fourth group containing particles of about 40 nanometers, each in a ratio of 62.8%, 16.2%, 16.7%, and 4.3%, and the aforementioned powder has a porosity of about 4.4%. In one embodiment, the aforementioned granular composite material powder is in a solvent. In one embodiment, the aforementioned granular composite material powder is suspended in a paste.

[0015] The present invention also considers the creation of layers using two or more different powders. In one embodiment, the present invention considers a method for producing layers comprising the steps of: a) providing first and second granular composite material powders, each having a porosity of 20% or less; b) depositing the first powder on the surface of a target object; c) accumulating energy in the first powder under conditions such that the energy causes sintering, fusion, or melting of the first powder to create a first layer; d) depositing the aforementioned second powder on the first layer; and e) accumulating energy in the aforementioned second powder, so that the energy causes sintering, fusion, or melting of the aforementioned second powder to create a second layer. Again, the present invention is not intended to limit itself to the energy source. In one embodiment, the aforementioned energy is accumulated by a laser. Again, the present invention is not intended to limit itself to layers of a specific thickness. In one embodiment, the aforementioned first and second layers are 100 microns (or less than 50 microns) thick. Again, the present invention is not intended to limit itself to any particular type or combination of particles used to produce the powder. In one embodiment, the aforementioned first powder comprises metal particles. In one embodiment, the aforementioned second powder comprises ceramic particles. In one embodiment, the aforementioned first powder comprises cermet particles. In one embodiment, the aforementioned second powder comprises a mixture of ceramic particles and metal particles. In one embodiment, the aforementioned first powder comprises titanium alloy particles and has a porosity of about 10%. In one embodiment, the aforementioned first powder comprises first and second particle groups, wherein the particles of the first group have an average particle volume at least 25 times greater than the average particle volume of the particles of the second group. In one embodiment, the aforementioned first powder comprises first and second particle groups, wherein the particles of the first group have an average particle volume between 25 and 2,000 times greater than the average particle volume of the particles of the second group.

[0016] The foregoing systems, layers, and methods can be used for 3D printing. For powders used in 3D printing employing an energy source for sintering, melting, or fusing of particles, a low porosity means, among several advantages, in particular, more reproducible manufacturing, high thermal conductivity, and high-efficiency energy absorption by the powder.

Brief Description of the Drawings

[0017] [Figure 1] FIG. 1A is a plot of volume probability density curves corresponding to two groups of particles, where Group 1 has an average volume V1 of 25 mm 3 of, V1 = 18 mm avg , a maximum volume at V1 = 18 mm 3 , and a standard deviation of its average (1 / 2) (blue), and Group 2 has an average volume V2 of 1,000 mm 3 of, V1 = 716 mm avg , a maximum volume at V1 = 716 mm 3 , and a standard deviation of its average (1 / 2) (green). FIG. 1B shows the passing curves corresponding to the same two groups of particles, plotted corresponding to S = 1.01005, assuming that each particle group contains the same total particle volume. To emphasize that the scale of the Y-axis in FIG. 1B depends on the parameter S, the Y-axis is named "particle fraction", and if each point V S,i is plotted and each point represents the particle volume between 0.5*(V s,i-1 +V S,i ) and 0.5*(V S,i +V S,i+1 ), then the unit of the Y-axis at that time is volume (mm 3). Note that when normalizing the integral value P(V) to 1, or when comparing P(V) for different particle groups in general, information about the relative number of particles in each group is not present in Figure 1A. However, the passage curve or probability density function for already combined composite materials will retain this information. Note that in the drawn graph, the positions of the maxima V1 and V2 are not the same, and the maxima appear on the passage curve (below figure) that appears at approximately V1 = 28.1 and V2 = 1113.2. Plotting the volume probability curve as a passage curve shifts the maxima and changes the height and standard deviation so that direct comparisons between particle groups can be made. [Figure 2] Figure 2 shows an example of a passage curve plotted for S = 1.00958, which includes five maxima (named V1 to V5). For this particle distribution, there are two "sufficiently large" particle groups corresponding to the maxima V2 and V5, with a volume range [V l, V 3-4 ], and [V 3-4, V r Each of these particle groups is represented by [ ]. These particle groups are also adjacent because there are no other sufficiently large particle groups between them, and the ratio of the average particle volume of each group to the smaller group is less than 10,000. The maximum value V1 is the smallest minimum value V between the maximum value V1 and V2. 1-2 However, the height of the maximum value V1 is PC s (V1) Larger height PC exceeding 75% s (V 2-3 Therefore, it is not associated with particle groups of sufficient size. The particle group corresponding to the maximum value V3 is the passing curve PC s (V S,i [V 2-3 ,V 3-4 The integral value of P(V)V over the range of ] (or the same range of PC s (V S,i The sum of ) is [V l ,V r The integral value of P(V)V over the range of ] (or the same range of PCs (V S,i Since it is less than 2% of the sum of (), it is not large enough. The local maximum V4 is similar to the local maximum V1, and the local minimum V 4-5 PC height s (V 4-5 ) However, the height of the maximum value V4 is PC s Since it is higher than 75% of (V4), it is not associated with a group of particles of sufficient size. [Figure 3] Figure 3 shows particles with a larger diameter of 1 and smaller diameter of 0.45 (a volume ratio of 11), and the relative volume fraction x of the smaller particles. s The pair correlation function g2(r) for a mixture of two spherical particles (of two sizes) is shown, where r is 0.267 (close to the minimum porosity relative to the particle size ratio). Note the peaks (discontinuities) and linear behavior in g2(r) at distances r=1.175, r=1.4, and r=1.725. These discontinuities, described as sharp maxima followed by a steep vertical drop, represent a high-probability linear alignment at the centers of contacting particles in a group consisting of two small particles and one large particle (r=1.175), one large particle between two small particles (r=1.4), and two large particles (r=1.725) and one small particle. The preference for these linear alignments at the centers of three particles is inherent to the DSMG structure, and each and all particle groups exhibit particularly small size variations in the mean size, i.e., the volume distribution P(V) for each independent particle group exhibits a small standard deviation. [Figure 4A] Figure 4A shows exemplary passage curves corresponding to grouping particles into two groups of sufficient size. [Figure 4B] Figure 4B shows an example of dividing particles into two groups of sufficient size. [Figure 5] Figure 5A shows an example of a passage curve corresponding to two groups of particles that should be considered as one, as there is no combination of relative particle volumes from Group 1 (blue, low maximum) and Group 2 (green, high maximum), and therefore there cannot be a combined passage curve for two sufficiently large particle groups. Figure 5B is an example of a combined passage curve for grouping of particles combined as a single group. [Figure 6] Figure 6 shows the group average particle volume V j avg This is a schematic diagram showing the separation of a 9-particle group j=1...9 into subsets that follow a volume range with a 4-order-of-magnitude fluctuation (10,000 times). In this case, the superset {i} of all subsets includes the four subsets {1,2,3,4}1, {4,5}2, {5,6,7}3, and {7,8,9}4. [Figure 7] Figure 7 shows side view (Figure 7A) and top view (Figure 7B) of a mixture with a void ratio of 21.2%, obtained by mixing soda lime beads with a diameter of 2 mm and a relative volume ratio of approximately 17%, with soda lime beads with a diameter of 10 mm and a relative volume ratio of approximately 83%. These beads have a sphericity of >0.98 and a static friction coefficient of <0.05, and therefore well simulate a frictionless sphere. [Figure 8] Figure 8 shows the approximate mixing volume fraction φ used to reduce porosity when mixing particle groups from different subsets. j This is a schematic diagram illustrating an example of volume calculation used in the calculation. This example is given for a subset that includes one particle group that overlaps with each adjacent subset. [Modes for carrying out the invention]

[0018] definition As used herein, the “TJ algorithm” refers to a method for producing irregularly tightly jammed (mechanically stable) packings of spherical or non-spherical objects (a packing is a collection of non-overlapping objects of specified location), as described in A.B. Hopkins, F. H. Stillinger, and S. Torquato, Disordered strictly jammed binary sphere packings attain an anomalously large range of densities, Physical Review E 88, 022205 (2013). The TJ algorithm tackles the problem of producing a strictly jammed packing as an optimization problem solved using linear programming techniques. The objective function to be minimized in this optimization problem is chosen to be approximately equal to a negative packing ratio, where the packing ratio is the spatial volume occupied by the objects.

[0019] The occupied space is a lattice vector M containing N objects under periodic boundary conditions. λ ={λ1;...;λ d This is a deformable unit cell in d dimensions accompanied by}. Each of the N objects consists of spheres of different sizes, so that the spheres overlap and maintain fixed positions relative to each other, but not for the other objects. Any object of any shape and size can be formed using overlapping spheres whose positions are fixed relative to each other.

[0020] Starting with initial conditions for the packing of N objects at an arbitrary packing density, and with irregular initial conditions at low packing densities, experimentally reproducible results are obtained, and the linear programming problem is solved to minimize the volume of the unit cell under the conditions of limited movement and unfolding of the N objects, shearing and compression of the cells, and linear arrangement and non-overlapping of the objects. This solution yields new coordinates and orientations corresponding to the objects, and new unit cells with smaller volumes. Using these new coordinates, orientations, and unit cells, a new linear programming problem is solved to minimize the unit cells under similarly limited movement and non-overlapping conditions. This process is repeated until further volume reduction in the unit cell is impossible. The final solution of this successive linear programming (SLP) process guarantees exact packing (mechanically stable).

[0021] The mathematical formulation of each linear programming problem is as follows: In this formula, r λ ij =x λ i -x λ j is position x λ i This is the displacement vector corresponding to the sphere, and the unit cell grid M λ Based on this, between the spheres i and j of the packing, △r λ ij =△x λ i -△x λ j ε = {ε} is the change in displacement solved between each SLP optimization step, and ε = {ε} kl} is similarly solved during that SLP step {ε kl This is the strain tensor associated with its unit cell, having}. λ i and ε kl The sphere between each step touches the upper and lower boundaries because it creates a limited range of movement that is small compared to its diameter.

[0022] The minimized value is the footprint of the strain matrix, Tr(ε)=ε 11 +...+εdd , and this corresponds to the linearized change in the volume of that unit cell. △x λ i and ε kl In addition to the upper and lower boundaries, each △x λ i and ε kl The spheres must be subject to linearized non-overlapping constraints for each pair of distinct spheres i and j. These constraints must be met by the average diameter D of spheres i and j. ij accompanied by, M λ ·r λ ij ·ε·M λ ·r λ ij +△r λ ij ·M T λ ·M λ ·r λ ij ≥(D 2 ij -r λ ij ·M T λ ·M λ ·r λ ij ) is described as follows. If spheres i and j are parts of the same object, then for each non-overlapping object, these spheres do not need to obey their linearized non-overlapping constraints. Rather, the positions of these spheres are fixed with respect to a single reference sphere, one reference sphere per object, and so the individual displacements of these spheres are entirely determined by the displacement and orientation of the reference sphere. For this purpose, for each reference sphere, two additional orientation variables φ i and θ i Therefore, φ for each reference sphere is necessary, and for that reason φ i and θ i is △x λ i and ε kl It must be tangent to the upper and lower boundaries and be solved corresponding to each SLP step.

[0023] Frictional resistance is also the "thickness" of the sphere, probability P f , 0≦P f ≤ 1, and distance xf This is incorporated into the TJ algorithm via the following: If two objects come into contact with each other at a specific point (i.e., within a distance x) in the next SLP step after a certain SLP step, they have a probability P through constraints on the movement and orientation of each object. f Maintain contact at a larger P f and x f A higher value indicates greater frictional resistance. Value P f =0 corresponds to the object having no friction or low frictional resistance.

[0024] In a two-phase heterogeneous medium (or "structure"), the dispersion σ of the partial volume fraction of each phase 2 i (R) is equal to (1 / v(R))*Int(χ(r)*α(r;R)dr), where "r" is a vector in d-dimensional Euclidean space, "dr" is an infinitesimal volume element in that space, v(R) is the volume of a sphere of radius "R" in "d" dimensions, χ(r) is the autocovariance function, α(r;R) is the scale cross volume, and the integral "Int" operates on the entire space. If the structure is very uniform, then the number variance σ i 2 (R) is (1 / R) d Rather (1 / R) d+1 It increases proportionally at a rate of . This is equivalent to saying that at the limit where ||k|| approaches zero, the spectral density, which is the Fourier transform F[χ(k)] of the autocovariance function, is equal to zero, where ||·|| represents the Euclidean distance and F[·] is its Fourier transform.

[0025] Consider a two-phase medium in which one phase consists of granular particles of any substance, type, size, or composition, and the other consists of voids, solvents, gels, pastes, or other types of fillers. For this medium, which is a granular composite, the partial volume fraction of phase "i" at point "z0" with respect to a given "R" is defined as the void fraction belonging to phase "i" contained within a d-dimensional sphere of radius "R" with point "z0" at its center. The dispersion σ of the partial volume fraction with respect to phase "i" is...i 2 (R) is the variance of the partial volume fractions of all points "z0" in the medium. The variance corresponding to such a two-phase medium does not depend on either phase and is expressed as a function of the autocovariance function χ(r) and the scale cross volume α(r;R) as σ in the mathematical description above. i 2 (R) is shown.

[0026] The autocovariance function χ(r) in a two-phase heterogeneous medium is given by χ(r) = S i 2(r)-φ i 2 The 2-point probability function S i Described as a function of 2(r) and the volume fraction φi, the same χ(r) is obtained regardless of which phase "i" is chosen. The 2-point probability function S corresponds to phase "i". i 2(r) is equal to the probability that both endpoints of a line segment of length ||r|| along the direction of "r" converge to phase "i", and the volume fraction φ i is the spatial fraction occupied by phase "i". The scale crossing volume α(r;R) is equal to the aggregate of two d-dimensional spheres of radius "R" separated by distance "r". For more details, see CEZachary, Y. Jiao, and S. Torquanto, Hyperuniform long-range correlations are a signature of disordered jammed hard-particle packagings, Physical Review Letter 106, 178001 (2011).

[0027] For a finite number of constituent particles, the limit of the spectral density F[χ(k)] at which ||k|| approaches zero must be defined in more detail, and ||k|| becomes zero only at the limit of infinite space and infinite particles. For a group of finite particles, the selected ||k|| value has a limit such that ||k||>1 / L, where "L" is the linear range in its system size. In this case, the limit at which ||k|| approaches zero can be defined by applying the curve of F[χ(k)] to some of the smallest ||k|| points and extrapolating the applied curve so that it includes the point ||k||=0. If the limit at which ||k|| approaches zero is the average effective diameter of its constituent particles... <V l / d >F[χ(k)] / <V l / d > 2 In that unit, if the value is less than "0.01", the structure is defined as "approximately very uniform".

[0028] This value "0.01" is chosen to reflect the degree of long-range spatial correlation present in a particular type of system. For example, all crystalline and quasicrystalline particles are highly uniform, as is the case with the most irregularly packed arrangement of spheres. All of these systems represent long-range arrangements between particles, where in crystals, the position of each particle is determined relative to its neighbors, and in the most irregularly packed arrangement of spheres, its pair correlation function is -1 / r, with "r" being the distance between points. d+lIt decays as follows (see A. Donev, F. Stillinger, and S. Torquato, Unexpected density fluctuations in jammed disordered sphere packagings, Physical Review Letter 95, 090604 (2005)). In liquids, the pair correlation function decays exponentially quickly, and at the limit where ||k|| approaches zero for a rigid sphere in a liquid, for example, F[χ(k)] is approximately equal to 0.028. Generally speaking, at the limit where ||k|| approaches zero, the smaller F[χ(k)], the greater the number variance σ in "R". i 2 (R) shows little increase, and spatial correlation between particles at long distances increases.

[0029] Description of the present invention This paper describes a density-enhancing process for granular composite materials, applicable to granular composite materials with constituent components of all sizes, shapes, and physical characteristics. This process consistently yields composite materials exhibiting lower porosity and viscosity combinations for a given particle known as a given particle within a range of relative sizes. Here, unless otherwise specified, particle size refers to the spatial volume occupied by a given single particle. This process also leads to a reduction in particle phase separation, where the separation of particles into different size groups is a significant practical obstacle in the production of high-density composite materials, especially when using multiple particle groups with large size differences between groups as components of a composite material. This process, which combines and mixes specially selected particle groups in desired ratios, produces previously unspecified types of compositions of structure and, consequently, material. These structures can be distinguished through their physical characteristics, including porosity and viscosity, and also, non-limitingly, through statistical measures, including structure-versus-correlation functions, contact distribution, and component volume distribution. Specifically, certain characteristics that appear in these structures according to the aforementioned statistical measures do not appear in the structures of well-known granular composite materials, even by those same statistical measures.

[0030] This process has numerous applications for desirable reductions in porosity, viscosity, phase separation tendency, or a combination of all three, and for desirable physical and related economic characteristics that occur as a result of and correlated with the reductions in porosity, viscosity, and phase separation. These desirable physical characteristics include, but are not limited to, a large bulk modulus, elastic modulus, shear modulus, durability, hardness, fluidity (ease of flow), thermal conductivity, heat capacity, electrical conductivity, overall absorption of laser (and other optical) energy, and overall absorption of heat (and optical energy), as well as a reduction in interfacial energy of "macro" molecules in solids, thermal expansion coefficient, depth from the skin for laser (and other optical) energy absorption, and depth from the skin for acoustic energy absorption. These desirable economic characteristics include, but are not limited to, a reduction in the cost of composite material components, increased reproducibility and repeatability of processing granular composite materials, and increased coherence of the processing. Applications in which these characteristics are beneficial include, but are not limited to, 1) additive manufacturing ("3D printing") involving metals, ceramics, cermets, polymers, plastics, or other dry or solvent-suspended powders, gels, or slurries; 2) concrete materials; 3) solid propellant materials; 4) cermet materials; 5) granular armor materials; 6) glass-metal and glass-plastic mixtures; and 7) ceramics including (or manufactured using) granular composite materials.

[0031] In additive manufacturing, the material, often in powder form, must be arranged in a desired spatial shape and react (usually by heating) with the underlying solid material, so that the primary material undergoes a phase change and combines with the solid, becoming solid itself. In these cases, the advantages of lower porosity and viscosity may, without limitation, include more uniform heating and melting of the granular composite material, easier and more uniform arrangement of the composite material, increased overall laser absorptive efficiency and reduced depth of absorption from the skin, reduced lateral energy scattering in the composite material, reduced oxidation of the composite material, and reduced temperature gradients in the entire composite material and the resulting melting and resolidification of the solid material.

[0032] In concrete materials, aggregate mixtures with lower porosity that still flow at the required rate can reduce the amount of paste needed to fill the gaps between aggregates and bond them together. These paste materials, generally including Portland cement, are often the most expensive component in concrete, and therefore their reduction is highly desirable. Furthermore, reducing the aggregate porosity in concrete often correlates with an exponential increase in strength, including bulk modulus, elastic and shear moduli, hardness, and lifespan.

[0033] In solid propellant materials, often granular composites, reducing porosity and associated increasing density leads to an increase in propellant thrust, which depends on the square of the composite density. Increased thrust is desirable for increased load launch capability, increased rocket speed, and other attractive advantages.

[0034] In ceramics, glass-metal mixtures, and glass-plastic mixtures, including (or manufactured using) cermet materials, granular armor materials, and granular composite materials, reduced porosity leads to greater strength, including, but not limited to, higher bulk modulus, elastic and shear modulus, hardness, and longer life. Reduced porosity also generally leads to increased durability under thermal stress cycles. These characteristics are usually desirable for these materials used as protective walls, load-bearing structural materials, and materials to withstand high temperatures, high stresses, and strains.

[0035] A. Characteristics of the novel composition The study of diverse mechanically stable mixtures of granular particles has led to the identification of novel compositions. These compositions exhibit characteristics similar to both liquid and solid compositions; for example, like a liquid, they will flow if a suitable external force (such as gravity) reshapes their equilibrium form, but like a solid, they will withstand small but non-zero volumetric and shear stresses without deformation. These compositions simultaneously exhibit solid and liquid characteristics, like a powder. These compositions are granular composite materials that can appear in powder form, slurry (liquid) form, or solid form if their components are fixed in place by a “paste” or “adhesive component,” as defined herein.

[0036] The process for producing this composition will be discussed later. This section discusses the defining characteristics and identification of this composition. This composition can be distinguished through its physical characteristics, including porosity and viscosity, and also, less limitingly, through statistical measures, including structure versus correlation function, contact distribution, and volume distribution. This section discusses certain characteristics that appear in these structures but do not appear in the structures of well-known granular composite materials, as described above.

[0037] This composition is defined by its structure and the composition of its underlying granular components. The structural classification including this composition will henceforth be referred to as the Dense Small-size-range Multimodal Granular (DSMG) structure. Certain structures within this classification are unusually dense for granular structures, considering the narrow size range occupied by these component particles.

[0038] In the following paragraphs, we assume that the monogranular composite material and the DSMG structure consist of at least approximately 100 particles. Generally, a DSMG structure can consist of far more than 100 particles; however, due to the irregular nature of the mixing of constituent particles in the monogranular composite material, at least approximately 100 particles are required for the characteristics of the DSMG structure to become clear.

[0039] All DAMG structures and mixtures of their constituent particles exhibit the following characteristics. 1. The DSMG structure consists of the sizes of the constituent particles that represent the volume probability density function P(V) (or simply the volume distribution, or particle size distribution), and includes two or more maxima associated with adjacent groups of particles of appropriate sizes. To define "adjacent" and "appropriate size" groups of particles, it is useful to first replot the volume distribution of the granular composite material as a particular type of passing curve that retains the maxima and minima (which will appear at slightly different positions) present in its volume distribution P(V), and then the granular composite material can be divided into groups of particles according to this passing curve. The replotting of the volume distribution as a passing curve is necessary to see the maxima and minima that appear at different volumes on the comparison scale. 1a. This volume distribution can be plotted as a set of points V S,i and is called the passing curve PC s (V S,i ) in accordance with the following method. For the volume range occupied between the volume of the smallest particle and the volume of the largest particle in the composite material, the volume intervals are selected on a geometric scale. In other words, each interval starts at X(i)=S i and ends at X(i + 1)=S i+1 , where "i" is an integer with i = m...n ("m" and [n] are probably either or both negative), and thus corresponding to some appropriately chosen scalar S>1, the "n"th power of "S" is greater than the volume of its largest particle, and the "m"th power of "S" is less than the volume of its smallest particle. The scalar S = S0 must be chosen to be at least sufficiently small so that for all S, 1 < S ≤ S0, the number of groups of particles of appropriate size determined by S and the associated maxima and minima always remain. This means that the volume distribution P(V) and the passing curve PC s (V S,iThis means that there is a one-to-one correspondence between the extreme values ​​(maximum and minimum) in ), and the extreme values ​​considered here are related to appropriate particle groups. As a result, P(V) and PC are related to appropriate particle groups. s (V S,i The corresponding maximum and minimum values ​​in both curves are considered to have the same meaning.

[0040] The given small "S" is the integral of P(V)V, which is obtained by multiplying the volume distribution by the volume, and is the result obtained by taking the values ​​at each interval and plotting them at the midpoint of each interval. That is, for each i=m...0...n-1, the integral value X(i)^X(i+l)P(V)VdV is the volume point V s,i It is plotted as =(1 / 2)*(X(i)+X(i+l)). To see the relative standard deviation (standard deviation divided by the mean) of the particle group on the comparison scale, point PC s (V S,i The curve ) must be plotted on a logarithmic volume axis. The resulting passage curve PC s (V S,i ) represents a geometrically measured mesh size, and therefore closely or exactly corresponds to the passage curve obtained by sieving particles using a standard sieve. s (V) The curve can be easily interpolated between points using any standard method, resulting in point PC s (V S,i It can be obtained from the set of ). Both curves PC s (V) and PC s (V S,i ) is a PC where i=m...n s (V S,i The sum of the ) is the average particle volume V avg Equal to V S,m From V S,n This represents the characteristic that the integral value P(V)V and its average particle volume are exactly the same. (Passage curve PC) s (V S,i Figure 1 shows a re-plot of the volume distribution P(V) for ). 1b. Once plotted, the passage curves can be easily divided into mutually exclusive, continuous volume ranges, and consequently, the particles of the granular composite material can be divided into size groups such that the endpoints of each volume range are the upper and lower boundaries of the particle size within that group. Particles at the endpoints where the volume drops can be placed in any of the groups that have those endpoints. The volume ranges are defined such that each represents a group of particles of "appropriate size," and the upper and lower boundaries of the ranges are at the minimums where the volume drops between the maximums associated with the groups of particles of appropriate size. An example and definition of "appropriate size" is as follows: Consider two maximums appearing in volumes V2 and V3 from the passage curves described in item 1a) above, where V3 > V2, and each of the maximums V2 and V3 corresponds to the average particle volume V2 of the group to which they belong. avg and V3 avg However, V3 avg / V2 avg The condition <= 10,000 is followed. This latter condition is one of two necessary conditions for particle groups considered to be "adjacent". For these maxima associated with appropriately sized particle groups, a) for all local minima appearing between the two maxima considered to be V 2-3 The size that appears on PC s (V 2-3 There must be a minimum value of "2-3" accompanied by ) and two maximum size PCs s (V2), PC s (V3) is approximately 75% of the smaller size, and b) a certain volume range [V l ,V r ] occupies V l <V2 avg <V3 avg <V r and V r / V l = 10,000, i) V S,i V1 and V2 and V 2-3 V between adjacent local maximums 1-2 The PC of all i is within the range of the minimum value that appears in S (V S,i The sum of ) and V S,i is range [V l ,V ris inside, and for all i within that range, PC S (V S,i ) the sum of which must be at least 2%, and ii) V S,i is V 2-3 and for all i within the range of the minimum value that appears between the maximum values adjacent to V3 and V4, PC 3-4 must be within the range of the minimum value that appears between the maximum values adjacent to V3 and V4, and for all i within that range, PC S (V S,i ) the sum of which is, for V S,i where V is within the range [V l , V r , for all i, PC S (V S,i ) the sum of which must be at least 2%. If there is no maximum value of V1 of a suitable size smaller than V2, then a volume V slightly smaller than the minimum particle size is used instead, and if there is no maximum value of V4 of a suitable size larger than V3, then a volume V slightly larger than the maximum particle size is used instead. If there are multiple minimum values between the paired maximum values, any one of the minimum values can be chosen, and it must be considered whether it meets criteria a) and b). After successfully applying to the maximum values of the granular composite material as a whole, it is important to note that criterion a) requires that both minimum values associated with the maximum values (one on each side) of the appropriately sized particle groups do not have a PC S (V) value greater than 75% of the maximum value between them.

[0041] The second condition for the two particle groups with the maximum values that appear at volume V2 and the "adjacent" V3 is that there must be no other maximum values at any V bt where the PC S (V) appears, and V2 < V bt < V3, so that for the pair of maximum values that appear at volume V2 and V bt and for the pair of maximum values that appear at V btFurthermore, the maximum values ​​of pairs appearing in V3 must satisfy criteria a) and b) that they are "sufficiently sized". Any granular composite material divided into particle groups according to the aforementioned criteria must all be separated into groups of sufficient size, and for the composite material to form a DSMG structure, it must contain at least one pair of adjacent sufficiently sized particle groups. The volume probability density function or passage curve to be referenced can be determined by sieving, centrifugation, image analysis, or any other common means. In particular, the function does not need to be precisely determined, provided that there is a normal level of accuracy obtainable by the common means of measuring them. Figure 2 includes an example of defining particle groups using passage curves.

[0042] 2. During mixing, spatial phase separation to particle groups related to particle maxima in the passage curve, according to the criteria described in Feature 1 above, cannot be applied to all groups across all spatial regions. Furthermore, all DSMG structures must exhibit some spatial mixing of particles from at least two adjacent groups of sufficient size. This means that for composite materials in their powder form, spatial mixing involves contact between particles of different particle groups in a spatial region of the composite material, where the height, width, and depth of that region are at least orders of magnitude within the lengths of multiple largest particles from a large volume particle group.

[0043] 3. When in powder form, that is, when there is no liquid or other medium substance to fill the gaps between constituent particles, the DSMG structure is in the group defined in Feature 1 above, group M i Particle G in i For a structure consisting of n groups, particle G i and G j At least one pair of the group (i and j are integers over all n, i is not equal to j) is in the group M i and M j In the mixing process, separate group G i and G jIt has the characteristic of creating a larger volume than the individual volumes of each particle group. Since the volume of a particle group depends on the preparation method, this characteristic generally requires similar methods for preparing and measuring the volume of a single particle group and a mixture of two particle groups. In particular, this characteristic excludes granular composites from the classification of DSMG structures that, in powder form, when all particle groups are mixed in the same amount as in their composite material, form a mechanically stable structure consisting only of particles from the larger volume particle group, where particles from the smaller volume particle group exist only in the voids created by the mechanically stable structure consisting only of particles from the larger particle group.

[0044] 4. Consider all sets of particle groups in the granular composite material, defined in Feature 1 above. In each set, the average particle volume of the largest particle group is not more than 10,000 times greater than the average particle volume of the smallest particle group. Taking these divisions into account, the DSMG structure requires the following: 4a. (When particle groups are ordered by the average particle volume of the group), it is possible to divide at least one set such that at least one pair of adjacent particle groups in the set is such that the volume of the "larger" group of the two groups is at least 20% of the total particle volume in the pair of particle groups, and the total particle volume of the "smaller" group is at least 10% of the total particle volume of the pair group. If these conditions are not met for all pairs of particle groups in all sets of particle groups, then it is impossible to satisfy these conditions, and generally, the granular composite material will not form a DSMG structure. 4b. For the entire composite material in powder form, considering all adjacent particle group pairs in all particle sets as described above, for at least one set, the average number of larger nearest neighbor particles to the particles in the larger group is 1 or greater. Volume V c For the target "central" particle, volume V nnThe "nearest neighboring particle" is defined as having a distance R s It is defined as a particle that can be positioned to come into contact with its central particle with a movement of less than R, where R S =((3V nn / 4π) 1 / 3 +(3V c / 4π) 1 / 3 ) / 2 is the mean spherical equivalent particle radius of the two particles. The nearest neighbor particle can be defined by a variety of techniques, including using an adhesive component or paste suitable for the purpose to solidify the powder structure in situ, performing image processing of cross-sectional slices of the resulting solid, and obtaining both the pair correlation function (obtained via standard scattering experiments) and the volume probability density function of the composite material through careful experimentation by those skilled in the art. The method for determining the nearest neighbor particle is independent of the DSMG structure, but a reasonable and accurate method is necessary to obtain reasonable and accurate results.

[0045] Feature 1, described above, describes DSMG granular composite materials in terms of particle groups of different sizes. Structures formed from a unimodal size distribution of particles (as described in Feature 1, a single particle group) are excluded from the classification of DSMG structures and reflect the requirement that a given particle group has the minimum total volume (relative to the desired volume range) in order to contribute to the formation of the structural features of DSMG. The ratio of the average size of larger particles in adjacent particle groups to the average size of smaller particles should not be too large. Feature 2 specifies a single structural feature possessed by a DSMG structure. Feature 3 describes a method for determining whether the structural features of DSMG are present in a granular composite material consisting of particle groups having specific physical characteristics, via a statistical measure of the constituent particles of a given structure. Feature 4 identifies the structural features exhibited by all DSMG structures and specifies a method for determining, using a statistical measure, whether the structural features of DSMG are present in a granular composite material consisting of particle groups having certain physical characteristics. Granular composite materials consisting only of sets of smaller and larger particles, where larger particles are dispersed to be diluted among smaller particles and rarely very close or in contact with each other, are excluded from the classification of DSMG structures.

[0046] In addition to these four features, all DSMG structures may exhibit one or more of the following features: A. A porosity of less than 25% for granular composite materials formed from only two particle groups, as defined in Feature 1 above. B. A porosity of less than 20% for granular composite materials formed from three or more particle groups, as defined in Feature 1 above. C. In powder form, without solvents, pastes, liquids, or gels between constituent particles, it is very uniform or nearly very uniform, as described in the Definitions section. D. An increase in the probability of at least one of three nearly linear arrangements at the center of a 3-particle ensemble consisting of two particles from one group and one particle from the other, between at least one pair of particle groups. The possible nearly linear arrangements are large-large-small, large-small-large, or large-small-small, corresponding to larger-sized particle groups and smaller-sized particle groups. For some composite materials, these ensembles can be detected by observing the pair correlation function of granular composite materials and can be obtained using scattering experiments. An example of detecting this feature, which is a depiction of the pair correlation function of a two-component mixture of spherical particles, is given in Figure 3. For composite materials consisting of particle groups described in Feature 1 above, the particles within individual groups will vary in size and shape, and the prominent peak in the probability of the pair correlation function at distance representing a linear ensemble of three particles consisting of two particles from one group and one particle from the other will be flattened and rounded. In many DSMG structures, this grouping may exist within the structure, but its features in the pair correlation function will be undetectable because they are flattened and rounded, as shown in Figure 3.

[0047] B. Production process of the composition We hereby describe a process for producing granular composite materials having reduced porosity, reduced phase separation tendency, and reduced viscosity, corresponding to particles within a specific ratio of the maximum volume of particles to the minimum volume of particles, in relation to composite materials known to be producible. The structure of the produced composite material often exhibits some or all of the structural and other physical characteristics described in the preceding section.

[0048] It is generally known that granular composite materials can be produced with a wide variety of porosities. For example, considering different arrangements of spherical particles of the same size, mechanically stable composite materials can be constructed with porosities ranging from approximately 51% to approximately 26%. In this usage, "mechanically stable" means that it can withstand a certain non-negligible strain or shear force from any direction without shrinking, deforming, or expanding, and therefore behave like a solid under a certain range of strain. However, structures that achieve the upper limit (51%) and lower limit (26%) mentioned are highly ordered in the sense that rigorous construction methods are required to produce structures that achieve these porosities.

[0049] Considering nearly spherical particles of the same size thoroughly mixed by vibration or shaking, and taking into account interactions via friction and gravity, a porosity of about 37% to 46% is common, depending on the exact shape of the particles and the coefficient of friction between them. This "thoroughly mixed" composite material is an important case to consider because the precise construction method is impractical, or even impossible, given the very large number of particles in most granular composite materials of interest. For example, even a handful of fine sand can contain more than a billion particles, and the time required to create a single-particle structure from such a handful at a time is considerable.

[0050] When considering granular composite materials composed of numerous particles of different sizes and shapes, the structure of a thoroughly mixed and mechanically stable composite material can range from a wide range of possible porosities. For example, experiments using particles with roughly log-normal size distributions produced porosities ranging from (approximately) 60% to 15%, depending on particle size, shape, and frictional interactions, along with detailed conditions of the composite material's formulation, such as whether the composite is vibrated or compressed at high frequencies. Since numerous manufacturing methods and natural processes produce composite materials with log-normal particle distributions, a log-normal particle distribution is important for practical applications. Examples of composite materials exhibiting roughly log-normal particle volumes include, non-limitingly, soil clods, mud, sand, several types of crushed stone, and several types of nanoparticles produced via flame or chemical methods.

[0051] Studies of concrete aggregate mixtures using cement, silica ash, sand, and crushed stone suggest that mixtures particularly suited to a roughly log-normal distribution of components produce a lower limit of 15% porosity. The most important factor for achieving the lowest achievable porosity is that the largest particles must have a far greater volume than the smallest particles. For example, in the referenced study, the largest particles have a volume more than 100 trillion times that of the smallest particles. (F. de Larrard, Concrete mixture proportioning, Routledge, New York (1999)).

[0052] In academic literature, hypothetical concepts of composite materials with nearly 0% porosity are discussed. Theoretically, this method is achieved by creating a continuous distribution of particle sizes, for example, a distribution of particle volumes with a very large standard deviation, so that the smallest particle has a minuscule volume compared to the largest particle, a difference of orders of magnitude between atoms and megaliths. In this method, if the size distribution is correctly chosen, the voids between larger particles can always be filled by smaller particles, and the porosity can approach 0% (although in reality it can never be reached).

[0053] Another theoretical approach discussed in academic literature is to use a discrete distribution of particle sizes to significantly reduce porosity, where the average volume of each smaller group of particles is small enough that smaller particles can easily occupy the space in the structure formed by larger particles. We call this the "large discrete ratio" approach, which reflects the idea that the composite material must consist of particle groups in which each group exhibits a discrete average volume at least 10,000 times larger or smaller than the average volume of all other particle groups in the composite material. An additional requirement for these groups is that the standard deviation of the particle volume distribution with respect to average volume must not be so large that the particle sizes of different groups substantially overlap. This 10,000-fold volume ratio reflects the minimum size difference required for smaller particles to fill the voids created by the mechanically stable structure of larger particles, and for small groups of smaller particles that do not form structures that hinder particle contact between larger particles. In "Disordered strictly jammed binary sphere packagings attain an anomalously large range of densities," Physical Review E 88, 022205 (2013), such a population is discussed for the case of frictionless two-component sphere composite materials.

[0054] When employing a large discrete ratio approach to create low-porosity composite materials, the particle volumes from each group must be selected in precise ratios such that each successive smaller group fits perfectly into the voids between the last larger group. For example, considering four groups of low-friction spherical particles, each group consists of particles of roughly the same size with a volume less than 100,000 times that of the next largest group. A mixture of 64.5% of the largest particle, 23.6% of the second largest, 8.6% of the third largest, and 3.2% of the smallest group (by volume) would produce a structure with a porosity of approximately 2.5%. In this example, it is noteworthy that the volume of the largest particle is 1,000 trillion times that of the smallest particle.

[0055] It is also noteworthy that, if the particle groups of a composite material produced using the large discrete ratio approach are vibrated or mixed after being placed in a gravitational environment, they tend to undergo phase separation, resulting in the smallest particle group at the bottom, the second smallest group above it, and finally the largest group at the top. This means that increasing the mixing or vibration time leads to a reduction in the porosity of the composite material, and in the aforementioned example using low-friction spherical particles with a small standard deviation of particle groups, the final porosity approaches 37%. This type of phase separation can, at best, sometimes be avoided if the ratio of the larger average particle volume to the smaller average particle volume between particle groups is as small as possible (for the large discrete ratio approach, the smallest such ratio is approximately 10,000). In such cases, one way to avoid phase separation is to create larger particle groups of material that are successively denser than the previous smaller groups. Another method involves continuously increasing the frictional interactions between particles, sometimes via electrostatic or van der Waals forces, but this method also generally increases the porosity of the composite material. Yet another way to mitigate phase separation is to compress the combined particle distribution from above, or to employ "upper-down" mixing rather than "left-right" mixing. Generally, the large discrete ratio approaches described here are not used in applications where a) the processing cost of particles with very similar sizes in each group is high, b) processing of particle groups is difficult when particles have volumes trillions of times larger or smaller than those in other groups, and c) overcoming particle phase separation is difficult.

[0056] It is not yet known to those skilled in the art or in public disclosures that another approach can continuously reduce the porosity of composite materials, and that this approach can be applied to any set of multiple particle groups exhibiting different size distributions, shapes, frictional interactions, or other physical characteristics. This approach does not require the largest possible differences in individual particle volume as in the aforementioned large discrete ratio approach. Furthermore, the composite materials produced by this approach do not readily undergo phase separation, unlike the composite materials produced by the aforementioned large discrete ratio approach, partly because the differences in particle volume between groups are not so large. This latter approach, which will be described in detail herein, is called the granular composite material density improvement process (GCDEP).

[0057] The GCDEP can be applied to any two or more given groups of individual granular composite materials, where the ratio of each group to show a representative passage curve for that group is intended to be combined in a specific proportion. Each group may consist of particles made from the same or different materials, and the shape and physical characteristics of the particles in a given group may be the same or different. In general, each particle group must show a single volume range in its passage curve that occupies where there is only one sufficiently large maximal, or where there is a maximal of roughly comparable size. If there are two or more maximals in a particle group, the two or more sufficiently large particle groups are identified in the manner described in Feature 1 (see above) for some combination of particle volumes from those sufficiently large groups, and the grouping must then be considered as two or more separated groups that conform to the separation criteria described in Feature 1. These groups can also be physically separated according to these criteria. This concept is illustrated in Figure 4. If a group cannot be physically divided, then the undivided group should be mixed with other groups, and if the volume ratio of particles is fixed among the sufficiently sized groups within that undivided group, then the undivided group should be considered as multiple sufficiently sized groups. Alternatively, if two or more particle groups cannot be divided into one or more sufficiently sized particle groups, and if the volume combination from the two groups cannot form two sufficiently sized groups in the volume range where their maxima or the location where their maxima overlap, then those groups should be considered as a single group. This concept is illustrated in Figure 5. When considered and prepared in this manner, every set of individual particle groups should be said to conform to the "GCDEP group criteria."

[0058] In this method, the GCDEP differs from other processes for reducing the porosity (and sometimes viscosity) of granular composites by considering the combined volume distribution of the entire granular composite rather than the mixing ratio of individual particle groups. Approaching mixing from the perspective of combined volume distribution is common to a number of fields, non-limitingly, including the formulation of high-performance concrete mixtures. For example, many methods discussed in academic literature, including F. de Larrard, Concrete mixture proportioning, Routledge, New York (1999) and JMSchilstone, Concrete mixture optimization, Concrete International, 33-40, June 1990, conclude that "gaps" in an uneven volume distribution can increase porosity and relative viscosity, whereas the GCDEP is based on the idea that a mixture formulation optimized based on gap size reduces porosity and relative viscosity. According to this and other descriptions, "relative viscosity" refers to the case where a granular composite material consists of particles suspended in a liquid or other type of matrix, where the relative viscosity is the viscosity of the particles and the matrix at a given particle ratio divided by the viscosity of the matrix alone.Other approaches, including the discussions in MN, Mangulkar and SS. Jamkar, "Review of particle packing theories used for concrete mix proportioning," International Journal of Scientific & Engineering Research, April, 143-148 (2013); F. de Larrard, "Concrete optimization with regard to packing density and rheology," 3rd RILEM international symposium on rheology of cement suspensions such as fresh concrete, France (2009); and F. de Larrard, "Concrete mixture proportioning," Routledge, New York (1999), sometimes consider different particle groups, precisely as the large discrete ratio approach takes into account. However, these approaches generally favor combinations of groups, and therefore the resulting composite materials do not conform to the maximum / minimum size criteria discussed in Feature 1 (see above).If these composite materials are undesirable, an assumption can be made, as in the case of the large discrete ratio approach, that smaller particles will occupy voids in mechanically stable structures formed by larger particles, as described in M. Kolonlo, S. Raschdorf, and D. Wasch, A hierarchical approach to simulate the packing density of particle mixtures on a computer, Granular Matter 12, 629-643 (2010). However, this assumption is inaccurate with large errors in the estimate of porosity when the average particle volume of the group is less than 10,000 in the ratio of large to small particles, and often results in mixed composite materials that readily undergo phase separation.

[0059] Some approaches that consider different groups attempt to correct their inaccuracies by assuming that larger particles will occupy voids within mechanically stable structures formed by larger particles. However, these approaches have two drawbacks, as discussed in F. de Larrard, *Concrete optimization with regard to packing density and rheology*, 3rd RILEM international symposium on rheology of cement suspensions such as fresh concrete, France (2009), and F. de Larrard, *Concrete mixture proportioning*, Routledge, New York (1999): 1) they cannot accurately predict the mixing ratio between particles that will produce the minimum porosity, and 2) they are not based on knowledge of the structures produced by the GCDEP and instead use, for example, inaccurate "virtual" structures. These “modified” approaches are fundamentally different from the GCDEP in question, and they represent formulations of mixtures different from those performed by the GCDEP. For example, these approaches most frequently require very large gaps between particles to produce a low-porosity structure, i.e., a large ratio of large average particle volume to small average particle volume. When the ratio of large average particle volume to small average particle volume is less than 10,000, these approaches support a particle combination, and therefore the resulting composite material's pass curve does not conform to the maximum / minimum size criterion described in Feature 1 (see above). Consequently, the predicted porosity and relative viscosity of the supported composite material will be much higher than those that can be achieved using the GCDEP.For example, as pointed out in "Self-consolidating concrete for precast structural applications: mixture proportions, workability, and early-age hardened properties," Center for Transportation Research, Project 0-5134-1 (2008), de Larrard's compression-filling model is known to overestimate the porosity of numerous granular composite materials, including composite materials consisting of two or more groups of roughly spherical particles. As discussed in ABYu, RPZou, and N. Standish, "Modifying the linear packing model for predicting the prosity of nonspherical particle mixtures," Ind.Eng.Chem.Res.35,3730-3741 (1996), this also applies to linear packing models and their extension to nonspherical particles.

[0060] Using the GCDEP, as illustrated herein, contrary to some other conclusions, “gaps” in the combined volume distribution of granular composite materials, when both the size of these gaps and the volume ratios of the individual groups mixed to form these composite materials are carefully selected, lead to much lower porosity and relative viscosity than gap-free materials. Examples using the GCDEP also demonstrate that, within the range of desired particle sizes, even greater reductions in porosity and relative viscosity can be achieved than those achievable by both 1) an approach that assumes smaller particles occupy voids in mechanically stable structures formed by larger particles, and 2) an approach that corrects the errors introduced by the assumption that smaller particles occupy voids in mechanically stable structures formed by larger particles. The GCDEP can achieve these reductions through a novel approach designed to advance the production of DSMG structures.

[0061] The volume probability density function for the constituent components of any granular composite material can be readily determined experimentally in a number of different ways. One method is the use of standard sieves, which is generally useful for many types of particles ranging in size from microns to centimeters. For smaller particles, their distribution can be determined using scanning electron microscopy and image analysis techniques, or using special sieves for small particles, or a benchtop centrifuge. Along with other methods not described herein, the sieving and centrifugation method can also be used to separate particles, and the separation (or combination) must conform to the aforementioned "GCDEP Group Criteria." The precise procedure used to determine the volume distribution is independent of the GCDEP in question, but is relevant insofar as increased accuracy in the procedure can reduce porosity and relative viscosity when using that process.

[0062] From this point forward, we will refer to materials made from combinations of different amounts of particles from two different groups as "two-peaked" composite materials, materials made from three different groups that meet the criteria as "three-peaked" composite materials, and four sets as "four-peaked" composite materials, and so on, as granular composite materials. This concept is most effectively applied in terms of reducing the porosity of granular composite materials containing combinations of different amounts of particles from individual groups when there is some overlap between any two volume probability density functions (or passage curves) of the particles of the individual groups. However, the GCDEP can be just as easily applied even when there is a large overlap between the individual distributions, which generally results in a higher porosity of the combined sets of particles.

[0063] The core of this is that the GCDEP can enable combinations of these groups, provided that the set contains two or more particle groups conforming to the GCDEP group criteria, in order to achieve low porosity and relative viscosity, including a method for calculating porosity, within experimental accuracy at a specific relative volume fraction. As previously used, "relative viscosity" refers to the relative viscosity of particles in a granular composite material, including the matrix material that fills the voids between the constituent components, where the ratio of constituent components to matrix material is fixed.

[0064] However, the GCDEP can further use a combination of optimization methods to select a subset of particle groups that minimizes or nearly minimizes porosity and relative viscosity from a set of three or more particle groups conforming to the GCDEP group criteria, and can provide all possible combinations of subset groups from that set. This section includes a description of the steps incorporated into the GCDEP and practical examples of their application to particle groups. The following sections include a description and examples of how the GCDEP is used to select a subset group from a set of three or more particle groups so that the porosity of all subsets relative to the particle groups is minimized or nearly minimized.

[0065] C. Improvement in the accuracy of the GCDEP When creating a granular composite material with a low porosity, before adopting the GCDEP, the available particle materials constituting the granular composite material can be identified. For example, when creating an ink used in the process of 3D printing, what initial materials will become the final printed member is generally known. Examples include pure components such as titanium, alloys such as stainless steel and Ti6-4, cermets such as titanium carbide or gallium nitride, and composite materials of mixed phases such as sand / aluminum. Once the details regarding what materials are available are established, a plan regarding their mixing method can be developed. It is important to note that the GCDEP does not prohibit any materials from being mixed in approximately the right proportions in order to obtain a reduced porosity and relative viscosity once the method is chosen.

[0066] Specific knowledge is useful for the GCDEP. This knowledge includes the approximate volume probability density function (or passing curve) of the selected group or groups of particles, and it is necessary to be able to confirm that the particle groups comply with the group criteria of the GCDEP. Alternatively, as described above, the particle groups can be separated or combined so that the final set of particles complies with the group criteria of the GCDEP. For example, in the case of a particle group showing a volume distribution reaching a wide range of particle sizes, this group can be divided into two or more different groups by excluding particles of a specific size, so that each new particle group is separated from other groups by a gap, and in its new set, the group complies with the group criteria of the GCDEP. It is important to note that when these steps are taken, the volume distribution becomes more accurate, the accuracy of the mixing ratio of the GCDEP increases more, and thus a lower possible porosity and relative viscosity can be achieved.

[0067] Rather than direct experiments, when using a lookup table or algorithm to complete step 1 of the GCDEP (discussed in more detail below), additional information is required in addition to the volume distribution of the particle groups. This information includes details about how the particles are mixed (e.g., slowly mixed under gravity or compressed, vibration, shaking, etc.) as well as details about the particle arrangement and frictional (e.g., electrostatic, van der Waals forces, etc.) interactions between the particles. For well-known materials, this information can be found in the relevant literature or an approximation of this information can be calculated. When applying the GCDEP, exact information is not necessary, but generally more accurate information leads to results that reduce the porosity and relative viscosity of the final particulate composite material developed using the GCDEP.

[0068] For a set containing three or more particle groups that conform to the group criteria of the GCDEP, prior to applying the GCDEP, a supersubset consisting of all subsets of the set of particle groups must be given in accordance with the following definitions of subsets and supersubsets. Subset {j} i is the range of volumes [V r / V l = 10,000, and consists of all particle groups having a group-average particle volume contained within that range [V l , V r , where {j} l , V r and where {j} i requests the "largest" subset "i", i.e., {j} i contains all the particle groups contained in {j} i and there is no subset of particle groups larger than {j} iThe superset {i} consists of the largest subset unique to all particle groups. This concept is illustrated in Figure 6. Thus, for a set consisting of three or more particle groups, the present invention considers criteria for dividing the set of particle groups into subsets that conform to the group criteria of the GCDEP.

[0069] Steps of D.GCDEP The steps of the GCDEP can be summarized in one embodiment as follows: 1. Determine the relative volume when particle groups within a subset are combined in a way that reduces porosity. This step can be achieved using computer simulations, experiments, or lookup tables. All three case examples are given below. 2. Taking into account the information determined in Step 1 for a subset of particles (a subset being clearly defined in the preceding section), calculate the relative volume of the particle group as a whole in that set, combined to reduce the porosity, in accordance with the formulas described herein, including the calculation of the approximate porosity to be achieved.

[0070] The important point is that if a particle group is formed from only a single subset, and the subsets do not overlap in terms of the particle group, then step 2 is not necessary. However, even for multiple subsets that overlap in terms of the particle group, it is possible to start (as step 1) with simulations, data tables, or experiments to find the minimum porosity structure rather than using step 2 of the GCDEP described above.

[0071] To refer to this step in more detail, the first step required by the GCDEP is to determine the relative volumes when the particle groups within each subset are mixed to minimize their porosity, for a set of particle groups conforming to the GCDEP group criteria. This can be achieved laboratory-wise using well-designed experiments, or lookup tables or computer simulations, assuming that their volume distribution, rough particle shape, and frictional interactions between particles are known. The use of one computer simulation to determine the minimum porosity of two particle groups consisting of a single size of frictionless spheres is described in the work of Hopkins, Stillinger, and Torquato (mentioned above). The computer simulation algorithm used in the work of Hopkins, Stillinger, and Torquato can also simulate three or more particle groups, and it can take into account non-spherical particles of arbitrary shapes and frictional interactions, although this increases the computation time by including groups of finer particles and highly aspherical shapes. This TJ algorithm is described in more detail in the Definitions section.

[0072] Using the volume probability density function of particle groups, in some cases, the relative group volume at which minimum porosity occurs can be roughly calculated with minimal error using this large discrete ratio approach. This generally applies to the following cases: 1) when the ratio of the average particle volume of the large particle group to the small particle group is at least 10,000, and 2) when, for a mixture of groups with relative volumes corresponding to minimum porosity, the individual volume distributions extend to a larger volume range than approximately 3% of the total volume of both particle groups, and the integral of the combined volume distribution of both groups does not exceed approximately 10% of the total volume of both particle groups. It is important to note that the figures 3% and 10% in these "overlapping integration guidelines" are not intended to be precise boundaries. Rather, the intention of this rough description is to make estimates based on existing data when two particle groups are processed using this large discrete ratio approach.

[0073] If the large discrete ratio approach can be adopted, the mixing of the two particle groups will not result in a structure that clearly exhibits some or all of the distinct features of the DSMG structure described earlier. However, the possibility of phase separation should be considered, as it occurs more frequently and more easily at larger ratios of average particle volume. In the case without phase separation, the total packing ratio φ in the large discrete ratio approach t (1 - equal to porosity) is the packing efficiency φ of a particle set containing only large volume particles. L and the packing density φ of a particle set containing only small volume particles S accompanied by, φ t =φ L +(1-φ L )*φ S This can be easily calculated. The relative particle volume x from a smaller set. S is, x S =((1-φ L )*φ S ) / φ t According to this, and the relative particle volume x from a larger set. L is, x L = 1 - x S =φ L / φ t Follow the rules.

[0074] For subsets consisting of three or more particle groups, the experiment can be used to find the relative volume of the particle group within the subset where the minimum porosity appears, although this experiment is time-consuming and laborious. We describe one experimental procedure that considers only two particle groups and determines the relative volume when the groups are mixed to obtain the minimum porosity. It is important to note that the GCDEP does not require the use of this procedure if any experimental procedure that can determine the minimum porosity is available. However, since the degree to which the porosity and relative viscosity of a granular composite material are reduced depends on the precision of the procedure adopted, it is recommended that the experiment be conducted very carefully to ensure precision.

[0075] Such a procedure requires a container capable of measuring volume with great accuracy and a precise scale. The high-precision container must be able to measure length, width (or diameter), and height at least 100 times the average longest linear range of the particles in the larger set. In addition, when measuring the particle volume, the container must contain enough particles so that the measured height is at least 100 times the average longest linear range of the particles in the larger group. This is due to the effect of the composite material's boundary walls, where larger particles can accumulate on the surface of the composite material by accumulating vast quantities of smaller particles in the voids of walls that larger particles cannot enter.

[0076] For example, consider two sets of particle groups, where the elongated, larger particles have an average volume 125 times greater than the smaller particles, and both sets have a standard deviation of less than 10% of their average volume. If, to obtain the minimum porosity, the materials are mixed in a volume ratio of 25% smaller and 75% larger in a container with a diameter and height 1000 times the longest linear range of the largest particles in the composite, and then, if, using a container with a diameter 10 times the longest linear range of the largest particles in the composite, the minimum porosity will increase, and the mixing ratio found to produce this porosity will be approximately 70% larger particles and 30% smaller particles. This rough approximation is based on the assumption that, if the boundary wall is within half the length of the average maximum range of the larger particles, the composite will contain 50% of the particle volume from the larger set and 50% from the smaller set. In fact, this is a very rough approximation, and the actual effect of the boundary is that, due to the spatial correlation between particles in mechanically stable granular composite materials, the effect is real even far from the boundary, beyond half the length of the average maximum range of larger particles. Furthermore, smaller and larger particle shapes, interparticle frictional interactions, and frictional interactions between particles and the boundary will affect the boundary effect, as will the way the particles are placed in the container. Therefore, the best way to reduce the relative effect of the boundary is to increase the surface-to-respect volume of the container by making the container larger and filling it with a larger particle volume.

[0077] Once a container has been selected, the zero-void mean density (density calculated by eliminating voids from the volume in the density denominator) of the individual particles in each set must be obtained. In particular, if the particle material is known and its characteristics can be found in the literature, there are numerous ways to achieve this. The method for determining the density is not important to the GCDEP in question, but a large error in density will lead to large porosity and relative viscosity results when the particle groups are mixed.

[0078] The porosity of each individual particle group should be determined using density data. This determination can be achieved in numerous ways, non-limitingly, including placing the container on a scale and, taking its mass into consideration, filling the container with particles from a single group to a height of at least 100 times the average longest linear range in that group. The difference in mass between a container filled with particles and an empty container is its packing efficiency (1 - its porosity), which is obtained by dividing the volume of particles in the container by the average density of the individual particles in that group. This measurement needs to be repeated many times and the results averaged to obtain high statistical accuracy.

[0079] When measuring volume and mass using scales and containers, utmost care must be taken to prepare the structure of each composite material in much the same manner as before. Since the preparation of the composite material strongly affects the porosity, this method must reflect the expected end use of the granular composite material. For example, if the granular composite material will be compressed in its end use, the preparation must include a compression step. If the granular composite material will be placed in a solvent in its end use, this must be added to the preparation (and the mass of the solvent must be taken into account and appropriately reflected in the calculations).

[0080] This approach needs to be repeated, except when considering the mixing of two particle groups. Again, care must be taken to ensure that the mixing of these two particle groups reflects the mixing that will appear in the final use of the composite material. One can start with only particles from the larger group and then gradually increase the mass of particles from the smaller group, or start with particles from the smaller group and gradually increase the mass of particles from the larger group. The volume fraction of the granular composite material needs to be calculated after each step of adding small amounts of particles to the mixture and after the step of remixing using the initial mixing method. Smaller mass increments will lead to a more accurate determination of the volume fraction at which the maximum density appears. In this case, the volume fraction is equal to the sum of the zero porosity of the particles from each group in the mass in the container, divided by the volume of the mixed composite material in the container. For each particle group, its zero porosity volume is calculated as the mass of particles in the container divided by the average zero porosity density of the particles in that group. This entire experiment needs to be repeated many times and the results averaged to obtain high statistical accuracy.

[0081] The packing efficiency curve is plotted as a function of the relative volume of particles from either the smaller or larger group, producing distinctly different maxima (where the relative volume at each point is calculated as the zero-porosity volume of particles from a single group in the mass of the container, divided by the sum of the zero-porosity volumes of both groups in the mass of the container). The composite material at and around this maxima has a structure that exhibits some or all of the features of the DSMG structure described above. Generally, mixtures in the range from about 10% or less of the absolute relative small particle group volume fraction to about 30-50% or more of the maximum exhibit some or all of the features of the DSMG structure. In those features, the exact distance in units of absolute relative small particle group volume fraction below or above the maximum can be detected depending on the particle size and shape, volume distribution, mixing method, and frictional interactions between particles.

[0082] If two or more particle groups are considered, the experimental procedure can be carried out using combinations of three or more particle groups, as described. This is very time-consuming, however, those skilled in the art can shorten the time by focusing on the relative volume fraction obtained in experiments with pairs of particle groups that are more likely to produce density maxima. In the case of three or more groups, density maxima greater than 1 may appear, and by identifying the maximum value among these maxima, the porosity and relative viscosity of the composite material as a whole can be reduced more significantly.

[0083] Once the relative volumes that produce the minimization of porosity are determined for all subsets, the relative volume fractions in that set need to be combined so that DSMG structures with low porosity and relative viscosity can be determined to a good approximation using a single technique. This technique is not a mandatory step, but it is the second step of the GCDEP in question, and the partitioning into subsets is not entirely necessary; this step simply allows for a quick and relatively accurate approximation for time savings. If there is only a single subset of the particle group, this second step is not necessary for all, and if there is no overlap between subsets, i.e., there are no two subsets containing the same particle group, this second step is mathematically simple (though very precise).

[0084] In this technique, the following notation is used: The superset {i} of all subsets, where the particle group with the highest average particle volume follows set i=1, and in all remaining subsets (except for the particle group that follows only set 1), the subset of the particle group with the highest average particle volume is set i=2, and so on. The subset of all particle groups contained in each set "i" is {j}. iHere, "j" is the particle group where the largest group is j=1, the next largest group is j=2, and so on. Therefore, if subset 2 contains the second, third, and fourth largest particle groups, then {2,3,4}2 is the notation for that set. The spatial occupancy of each group in a subset is found from step 1 of the GCDEP using mixed experiments, simulations, lookup tables, or other means, φ i,j And, according to the void, each subset of particles "i" has a spatial occupancy rate φ due to the void. iv and subset {j} i The sum of all "j" in 1 = Σ ji φ i,j +φ iv This is the φ adopted in the second step of the GCDEP. i,j φ appears at the minimum value in the porosity corresponding to the mixing of subset "i". i,j It is important to note that this does not necessarily have to be the case. However, as a technique to improve accuracy, φ is used for each subset "i". i,j This needs to be positioned on the actual mixing curve (or surface) determined in the first step of the GCDEP. As discussed previously, the DSMG structure produces the minimum porosity φ i,j φ away from the value i,j A range of values ​​can be formed, and the second step of the GCDEP generally applies to this range of values.

[0085] A qualitative overview of this technology is as follows: When considering combinations of pairs of overlapping subsets between particle groups, some particle groups will "interact" because their sizes are very similar, while particles from the smallest subset (minimum particle group) will behave as if particles from the largest subset (maximum particle group) were the boundary of a limited void, i.e., as if these groups were "non-interacting". With this in mind, there is a volume of free space filled only by particles from the smallest particle group, bounded by the largest particle group. In the following notation, these volumes ending in the subscript "1" are referred to as subscript {α i For} and subset "i", their volumes are V{α1,α2,...,α (i-2) ,α (i-l) It is written as ,1}.

[0086] The volume space figures for each subset related to the terminology in the calculation of the final volume fraction of a particle group follow the Fibonacci sequence. The calculation of the first subset requires only one volume, represented by V0 (V0 must be equal to the volume of the filled space, and the volume fraction term is written as V0=1). The second subset is V 00 and V 01The first subset requires two volumes, the third subset requires three volumes, the fourth subset requires five volumes, the fifth requires eight, and so on, following the Fibonacci sequence. In subscript notation of two variables, the subscript number indicates the subset number "i", the term "0" indicates the volume possible (and occupied) by all particle groups in subset "i", and the term "1" indicates the volume possible only by the smallest particle group in subset "i", i.e., the group belonging to "i" and not subset "i-1". Furthermore, the subscript "1" cannot be adjacent to any other subscript "1", and a value possible only for the smallest particles in a subset cannot include an "additional" value for the next subset "i+1", which is possible only for the smallest particles in that subset. Following these rules in subscripts, the number of volumes in subset "i" corresponds to the Fibonacci number "i+1". For example, its fourth subset has five volumes V 0000 , V 0100 , V 0010 , V 0101 , and V 0001 This includes and "5" is the fifth Fibonacci number. Figure 8 illustrates the calculation of these volumes for the subsets {1,2}1, {2,3}2, {3,4,5}3, {5,6,7,8}4, and {8,9}5.

[0087] In contrast to this technology, single particle group (non-mixed) P j The porosity of may be required. Furthermore, in certain cases for specific subsets containing three or more particle groups, information is needed that takes into account the mixing of particle groups from larger groups or from subsets excluding larger groups. This requirement is described in step 3 below. In the techniques described below, quantity φ jrepresents the volume ratio of the total space occupied by the particles from the target particle group in the mixture of the final composite material. This technique tends to be slightly more accurate when each consecutive pair of subsets either has no particle groups or shares only a single particle group between them. Therefore, these cases are discussed first, and then the cases where any two subsets share multiple particle groups, which will be described later as special cases, are discussed.

[0088] This technique is as follows. 1) Start with the first subset i = 1 and assume that for all "j" of subset 1, φ j is equal to φ 1j . 2) When particle groups are shared between subsets, calculate the volume V{α n} of all the next subsets, that is, V 00 and V 01 of the second subset i = 2. a. For any volume with the last two subscripts "α (i-1) ,α i " ending with "0,0", V{α1,α2,...α (i-2) ,0,0}=V{α1,α2,...α (i-2) ,0}*(φ (i-1),j / φ i,j ), where "j" refers to the particle group shared by both subsets "i" and "i - 1". For example, if i = 4 for adjacent subsets {4,5,6}3 and {6,7,8,9}4, then V 0000 =V 000 *(φ 3,6 / φ 4,6 ), and V 0100 =V 010 *(φ 3,6 / φ 4,6 ). If V{α1,α2,...α (i-2) ,0,0} is greater than V{α1,α2,...α (i-2) ,0}, then V′{α1,α2,...α (i-2) ,0,0}=V{α1,α2,...α (i-2) ,0} and V′{α1,α2,...α(i-2) ,0}=V′{α1,α2,...α (i-2) ,0,0} / V{α1,α2,...α (i-2) Calculate V{α1,α2,...α (i-2) ,0} and V{α1,α2,...α (i-2) Instead of ,0,0}, V′{α1,α2,...α (i-2) ,0} and V′{α1,α2,...α (i-2) Using ,0,0} respectively, and in all subsequent calculations, V{α1,α2,...α (i-2) Any other volume requiring ,0} is instead V′{α1,α2,...α (i-2) Calculate using},0}. b. The last two subscripts ending in "1,0" are "α (i-1) ,α i For any volume of ", V{α1,α2,...α (i-2) ,1,0}=V{α1,α2,...α (i-2) Assume that there are no particle groups belonging solely to subset "i-1". If there were, then the void ratio occupied by the mixture of minimum porosity of all particle groups belonging to subset "i-1" but not to subset "i" must be known. These ratios are φ i’,j Described as, and V{α1,α2,...α (i-2) ,1,0} is V{α1,α2,...α (i-2) ,1,0}=V{α1,α2,...α (i-2) ,1}*(φ (i-1)′,j / φ i,j It is used to calculate as ), where "j" refers to the group of particles shared by both subsets "i" and "i-1". For example, for subsets {2,3,4}2, {4,5,6}3 and {6,7,8,9}4, if "i=4", then V 0010 =V 001 *(φ 3′,6 / φ 4,6 ) and here, φ 3′,6 This represents the volume fraction of group 6 that minimizes the porosity of the mixture of particle groups 5 and 6. c. The last two subscripts ending in "0,1" are "α(i-1) ,α i For any volume of ", V{α1,α2,...α (i-2) ,0,1}=V{α1,α2,...α (i-2) ,0}*(1-Σφ (i-1),j )-V{α1,α2,...α (i-2) ,0,0}, where φ (i-1),j The sum of all φ of particle group "j" that belongs to subset "i-1" but not to subset "i" is equal to all φ (i-1),j This includes. For example, for subsets {2,3,4}2, {4,5,6}3 and {6,7,8,9}4, if "i=4", then V 0101 =V 010 *(1-φ 3,4 -φ 3,5 )-V 0100 That is the case. 3) Any negative volume V{α1, α2, ... α (i-1) For ,1} as well, φ found in the first step of the GCDEP i,j Adjustments are needed. Two possible interpretations are: the first group generally has a slightly lower overall porosity than the second group, and the second group generally exhibits relatively less phase separation. a.φ′ (i-1),(j=x) against φ (i-1),(j=x) To reduce, here, φ′ (i-1),(j=x) The minimum volume V{α1,α2,...α} of subset "i" is (i-1) φ is defined as a value where φ = 1 is zero, and here the particle group "j = x" belongs to both subsets "i" and "i-1". (i-1),j Instead of φ′ (i-1),j Using this, all volumes V{α} for subset "i" n Recalculate} and all subsequent calculations φ (i-1),j Instead of φ′ (i-1),j Proceed using [this method]. b. This applies to the volume calculation of subset "i", but not to other volume calculations, φ′ (i-1),j against φ (i-1),j We reduce it proportionally, and here φ′ (i-1),j The minimum volume V{α1,α2,...α (i-1)φ is defined as a value where {1} is zero, and here particle group "j" is all groups belonging to subset "i-1" and not subset "j". (i-1),j Instead of φ′ (i-1),j Using all volumes V{α} of subset "i", n Recalculate}. If this step is adopted, then φ for subsets "i-1" and "i-2". i When calculating this reduced φ′ (i-1),j You need to use it. 4) Repeat steps 2) and 3) until the volume is calculated for all subsets. 5) Its volume V{α n Using}, φ for each particle group "j" j Calculate. In the following equation, φ′ discussed in step 3 i,j If instructed, φ i,j This can be replaced with: For each subset "i" starting with "i=1", a group of particles "j" smaller than subset "i" and not belonging to any "i" is considered. a. For a particle group of subset "i" that does not have a particle group belonging only to subset "i", φ j =φ i,j *ΣV{α1,α2,...,α (i-1) ,0}+φ (i+1),j *ΣV{α1,α2,...,α (i-1) ,1}, where the first sum is all V{α1,α2,...,α} of the subset "i" that does not end with the subscript "1". (i-1) ,α i The sum of} and the second sum is all V{α1, α2, ..., α} ending in the subscript "1". (i-1) ,α i It is the sum of}. If there is no subset of subscripts greater than "i", then (1-P j The value of the second term in the equation is φ (i+1),j It is replaced by φ7 = φ 4,7 *(V 0000+V 0100 +V 0010 )+φ 5,7 *(V 0001 +V 0101 ) . If, in the example above, subset 5 is not mixed, then φ7 = φ 4,7 *(V 0000 +V 0100 +V 0010 )+(1-P7)*(V 0001 +V 0101 ) b. For a particle group in subset "i" that has a particle group "j" belonging only to subset "i", φ j =φ i,j *ΣV{α1,α2,...,α (i-1) ,0}+φ i′,j *ΣV{α1,α2,...,α (i-1) ,1}, and here, its φ i′,j This is the same as the void fraction calculated in step 2b), i.e., the void fraction occupied by the mixture of minimum void fractions of all particle groups belonging to subset "i-1" rather than subset "i". For example, for subsets {2,3,4}2, {4,5,6}3, and {6,7,8,9}4 where i=4, φ7=φ 4,7 *(V 0000 +V 0100 +V 0010 )+φ 4′,7 *(V 0001 +V 0101 ), φ8=φ 4,8 *(V 0000 +V 0100 +V 0010 )+φ 4′,8 *(V 0001 +V 0101 ), and φ9=φ 4,9 *(V 0000 +V 0100 +V 0010 )+φ 4′,9 *(V 0001 +V 0101 φ7, φ8, and φ9 are calculated so that ) 6) If no particle groups are shared between subsets, the superset of all subsets can be subdivided into sets of subsets such that each subset shares a particle group with at least one other particle group in that set. For example, {1,2}1,{2,3}2,{4,5,6}3,{6,7}4,{8}5 can be subdivided into three sets: {1,2}1,{2,3}2;{4,5,6}1,{6,7}2; and {8}1, where each subset "i" has its own volume (V1){α n}, (V2){α n}, (V3){α n} is calculated in the way described through step 1) to 6) above. The only difference is that when (V1)0=1, for example (V2)0 is set to be equal to the total porosity of the particle group belonging to set 1, and (V3)0 is set to be equal to the total porosity of the particle groups belonging to sets 1 and 2, etc. Using the above example, (V2)0=1-φ1-φ2-φ3, and (V3)0=1-φ1-φ2-φ3-φ4-φ5-φ6-φ7=V20-φ4-φ5-φ6-φ7.

[0089] If there are more than two particle groups overlapping between subsets, modifications to steps 2a and 2b of the aforementioned six-step technique are necessary. This modification is simple: when calculating the volume of subset "i", (φ (i-1),j / φ i,j )(Step 2a) or (φ (i-1)′,j / φ i,j Instead of multiplying by (step 2b), multiply all overlapping group volume ratios φ i,j and φ (i-1),j It is necessary to use general reduction factors that take this into account. For example, for groups j1 and j2 that overlap between subsets "i" and "i-1", the average volume ratio of the overlapping group is (φ (i-1),j1 +φ (i,j),j2 / φ i,j1 +φ i,j2 )(Step 2a) or (φ (i-1)′,j1 +φ (i,j)′,j2 / φ i,j1 +φ i,j2)(Step 2b) is the original factor (φ (i-1),j / φ i,j )(Step 2a) or (φ (i-1)′,j / φ i,j )(step 2b) can be adopted in place of each. This average can be similarly extended to three or more overlapping groups, or if the larger average volume fraction of two overlapping groups is, for example, larger than 1,000 compared to the smaller average volume fraction, the fact that they overlap can be ignored, and for the purpose of volume calculation, it can be assumed that the group with the larger average particle size does not belong to subset "i". For example, for the target groups {1,2,3}1 and {2,3,4}2, its factor (φ 1,3 / φ 2,3 ) to, V 00 It can be used in the calculation. Ignoring overlaps in this case will result in a slight decrease in the accuracy of the final calculated porosity.

[0090] For all other steps in the aforementioned 6-step technique, any language applied to a single overlapping group can be extended to apply to all overlapping groups. For example, if we use the average volume ratio to calculate the reduction factor, in step 2c, for subsets {3,4}2, {4,5,6}3, and {5,6,7}4 where i=4, groups 5 and 6 that belong to both subset 3 and subset 4 are V 0101 =V 010 *(1-φ3,4-φ 3,5 )-V 0100 Rather V 0101 =V 010 *(1-φ3,4)-V 0100 However, if we assume that only group 6 is duplicated, then, as mentioned above, V 0101 =V 010 *(1-φ3,4-φ 3,5 )-V 0100 That is the case.

[0091] E. Method for selecting particle groups that minimize porosity In many practical environments, materials capable of forming particles can be divided into numerous groups, or there may be numerous particle groups from which composite materials can be formed. For example, in concrete manufacturing, there are often 10 to 50 different types of aggregates in conveniently located quarries or other aggregate production facilities near the site. Alternatively, in additive manufacturing, powder materials of specific molecular compositions that meet specific average particle volume and standard deviation criteria can be manufactured (or ordered). Generally, if numerous particle groups with average volumes within a small ratio range are available, for example, if there are 10 available groups with a maximum average volume ratio of 1,000,000, using all available groups to make up a granular composite material may not produce the minimum porosity. In these cases, the correct group selection is crucial to achieving low porosity.

[0092] Some "rules of thumb" for selecting particle groups are: a) choose adjacent groups where the average volume ratio of large particles to the average volume of small particles is greater than 25 but less than 2,000; b) choose groups with smaller relative standard deviations where possible (e.g., if the average volume ratio of large particles to the average volume of small particles is small for adjacent particle groups, choose groups where the arithmetic standard deviation is less than 30% of the average particle volume); and c) choose groups with high average sphericity and, if the mixing method does not include compression or vibration steps, choose groups with a low static friction coefficient.

[0093] Criterion 1(a) reflects a balance between reducing porosity and reducing the tendency for phase separation. When mixing different particle groups, both porosity and the tendency for phase separation depend heavily on a variety of factors. However, adjacent particle groups where the average volume ratio of large particles to the average volume of small particles is relatively small are far less likely to undergo phase separation but tend not to form a high-density structure, while adjacent particle groups where the average volume ratio of large particles to the average volume of small particles is relatively large are more likely to undergo phase separation but tend to form a high-density structure if phase separation does not occur. Criterion 2(b) reflects the differences between continuous particle distributions and adjacent groups of sufficient size. If particle groups are not clearly defined as described by the GCDEP criteria, these groups tend to result in a non-high-density packing when mixed. Criterion 3(c) reflects general guidelines for mixing multiple particle groups. Groups of high-angle, aspherical particles require both a relatively large average volume ratio of large particles to the average volume of small particles and careful formulation to achieve a composite material with reduced porosity. Achieving a low porosity is simpler when using spherical particles with a low coefficient of static friction.

[0094] Balancing the porosity of granular composite materials with other desirable physical properties, including, but not limited to, relative viscosity and phase separation tendency, requires considerable knowledge of the physical characteristics of particle groups, physical interactions, size distribution, particle geometry, mixing methods, and other potential factors. However, sufficient knowledge of some or all of these important factors, and the selection of particle groups to mix to achieve desirable results (e.g., finding a minimal porosity structure that does not undergo phase separation in a particular application given a modified upper limit on relative viscosity), can be obtained by considering particle mixing as an optimization problem.

[0095] Given a superset of particle groups conforming to the GCDEP group criteria, the first step of the GCDEP can be used in combination with an optimal method for minimizing the porosity of a granular composite material consisting of a subset of those particle groups. This approach is supported by further subdividing the superset of particle groups into subsets, as illustrated in the example in Figure 6, and by knowledge of the porosity of the subsets as a function of the relative volume of each group in the mixture, mixed according to the intended use of the granular composite material. i For the target subset [i] containing the particle group "j", their porosity is given by the volume fraction φ i,jn P i ( i,j1 ,φ i,j2 ...φ i、jN It can be written as ).

[0096] These porosity functions P i ( i,j1 ,φ i,j2 ...φ i、jN Given ), the group selection problem is Σ i (N i This becomes a nonlinear programming problem with a large number of variables equal to -1), where the sum is the sum of all subsets [i] and the term "minus 1" is the porosity P i and φ i,jn N i This is due to the fact that -1" was given, and its last φ i,jn This can be determined. These methods, familiar to those skilled in the art, can construct the objective function of a programming problem using the aforementioned six-step approximation technique or any other precise subset combination devised. Standard constraints may be included in this problem, for example, φ i,jn >=0, or to reduce phase separation, φ corresponding to the minimum porosity at any quantity. i,jn Any φ greater than any i,jnThe constraints can also be set by incorporating step 3 of the 6-step approximation technique. Any precise programming technique, including but not limited to the extended Lagrangian function, quasi-Newton method, barrier function, and conjugate gradient method, can be used, either by itself or in combination with other methods to obtain a solution with reduced porosity. In using these methods, for a large number of particle groups, there are always a large number of volume fractions φ when j=jn. i,jn The fact that it is zero indicates that group "j" is unlikely to be included in the final composite material. This non-zero volume fraction φ i,jn This indicates the group that should be included in the final composite material.

[0097] φ of each subset i,jn P across the scope of interest in i ( i,j1 ,φ i,j2 ...φ i、jN Obtaining partial knowledge of ) requires a time-consuming method. However, this knowledge is generally not necessary for the optimization method in question. Rather, the partial subset porosity function P corresponds to the mixing of particle group pairs such that the ratio of the average volume of the large particle to the average volume of the small particle in that pair is less than 10,000. j1,j2 ( i,j1 ,φ i,j2 ), and its function P for a mixture of three particle groups such that the ratio of the average volume of large particles to the average volume of small particles is less than 10,000. j1,j2、j3 ( i,j1 ,φ i,j2 ,φ i,j3 To obtain ), it is more practical if the number of particle groups in subset[i] is large. Particle groups such as 4, 5, and 6 are also adopted under the same volume ratio standard. However, if a particularly large superset of particle groups is given, the “empirical rule” mentioned earlier can be used to reduce the number of desirable porosity functions.

[0098] As shown in Figure 6, the porosity function P of a partial subset is shown. j1,j2 ,…(φ i,j1 ,φ i,j2Given the separation procedure for the ,…) and its subsets, a person skilled in the art can construct an objective function corresponding to the programming problem using the six-step approximation technique described above. The constraints are the overall function P i ( i,j1 ,φ i,j2 ...φ i、jN It can be added in the same way as when using ). Furthermore, with this approach, P j1,j2 ,…(φ i,j1 ,φ i,j2 ,…) but form a subset using the subset separation procedure explained in Figure 6, its P j1,j2 ,…(φ i,j1 ,φ i,j2 The use of the objective function of particles is required in groups corresponding only to the set of ,…). As a result, each of P j1,j2 ,…(φ i,j1 ,φ i,j2 It will be necessary to consider and solve several different programming problems, each with objective functions and constraints corresponding to different sets of P. j1,j2 ,…(φ i,j1 ,φ i,j2 Because of its simple form, it would be significantly more time-efficient in solving the variable reduction problem in each target programming issue.

[0099] Ultimately, the complete or partial porosity function P i ( i,j1 ,φ i,j2 ...φ i、jN ) and P j1,j2 ,…(φ i,j1 ,φ i,j2 ,…) are not necessary for selecting from a large superset of groups to mix in order to achieve a granular composite material with significantly reduced porosity. Each porosity function P i ( i,j1 ,φ i,j2 ...φ i、jN ) A single point from, or a selected set of P j1,j2 ,…(φ i,j1 ,φ i,j2Given a single point from each of ,…), where P j1,j2 ,…(φ i,j1 ,φ i,j2 Each of the sets of ,…) forms a subset using the subset separation procedure described in Figure 6, and the six-step procedure described in subsection B of the previous section can be used to obtain a mixture with reduced porosity. In particular, if a given point is the minimum or near minimum porosity of its porosity function, the six-step procedure described in subsection B of the previous section will produce a mixture with significantly reduced porosity.

[0100] Description of a Preferred Embodiment As described above, in one embodiment, the present invention considers additive manufacturing, non-limitingly, including selective laser sintering (SLS). SLS is a technique used for manufacturing prototype models and functional parts. SLS uses a laser as a power source to sinter powder material, bonding the powder particles together to create a solid structure. The laser beam is aimed at scanning the powder layer and switching to sinter the powder or a portion of the powder. A continuous sintered layer is formed as the powder is supplied and becomes a complete part. The powder may include plastics, metals, ceramics, carbides, glass, polymer materials (as well as combinations thereof).

[0101] Selective laser sintering and direct metal laser sintering (DMLS) are essentially the same method, but SLS refers to a process applied to a variety of materials—plastics, glass, and ceramics—while DMLS often refers to a process applied to alloys. However, the fact that it is a sintering apparatus different from melting or "fusion" means that the sintering process does not completely melt the powder, but heats it to a point where the powder can fuse together at the molecular level.

[0102] On the other hand, selective laser melting, or SLM (often called direct metal laser melting, or DMLM), uses a laser to achieve complete melting. In this case, the powders not only fuse together but substantially melt to form a homogeneous part. Melting is generally beneficial for a single material (pure titanium or a single alloy such as Ti6-4). In contrast, when working with multiple materials, alloys, or combinations of alloys with other materials such as plastics, ceramics, polymers, carbides, or glass, it is now common to use SLS or DMLS.

[0103] This invention considers SLS and DMLS using powders with increased porosity. Furthermore, such powders are expected to correspond to SLM and DMLM.

[0104] In general, in the field of additive manufacturing using composite materials such as powders, most achievements are not made with the expectation of optimizing the structure of the composite material. Even among experts in this field, it is not known that composite materials with very low porosity can be produced, and that these composite materials do not easily undergo phase separation and remain malleable enough for use in standard applications. Metal and ceramic powders used in selective laser melting or selective laser sintering processes generally exhibit porosity of about 30% to 50%. In the few available academic literatures on the effect of powder porosity in laser sintering and laser melting processes, a minimum porosity of 25% has been achieved. EOOlakanmi, KWDalrarno, and RFCochrane, Laser sintering of blended Al-Si powders, Rapid Prototyping Journal 18, 109-119 (2012). The lack of progress in low-porosity powders is, non-limiting, partly because a) the extent of the advantages of using powders with less than 25% porosity is unknown, and b) a low-cost production process is unknown for producing powders that do not easily separate phases, maintain malleability for easy processing (i.e., have sufficiently low viscosity), and exhibit porosity of less than approximately 25%.

[0105] This paper describes the use and benefits of low-porosity powders, defined as powders with a porosity of 20% or less, as a common process improvement in additive manufacturing processes using metals, ceramics, cermets, glass, carbides, or other high-melting-point powders that are concentrated and melted by lasers or other processes. These low-porosity powders can be produced by either increasing the density of granular composite materials or by several other processes (this distinction will be discussed in detail later). The use of low-porosity powders can begin with powder particles consisting of metals, cermets, ceramics, glass, polymers, and alloys of any elemental or molecular composition. In particular, the 20% porosity threshold is provided for low-porosity powders, reflecting the upper limit of porosity of powders used and available in current additive manufacturing processes.

[0106] In additive manufacturing processes conducted by industry and research organizations, several advantages of reducing the porosity of powders have sometimes been recognized. The “compression step,” sometimes employed in laser melting and laser sintering processes using rollers or vibrating hoppers, is clearly intended to reduce the porosity of the powder from approximately 35–50% to approximately 25–30% before melting or sintering using laser energy. However, in some cases studied to date, reducing the porosity beyond these values ​​is often undesirable. For example, one study containing data on the effects of reducing powder porosity in the laser sintering process of laser-sintered parts states that “results from [other] studies have not been able to fully determine what direct effects the characteristics of particle packing arrangement have on process conditions, densification, and microstructure of laser-sintered components.” See Olakanmi et al. (previously mentioned). Olakanmi et al. studied metal powders with porosities ranging from approximately 28% to 37%. While some of their low-porosity powders produced the desired very high-density printed parts, in other cases they found "no strong correlation" between the porosity of the printed parts and the powder. They concluded that it is unclear whether porosity reduction beyond the experimental parameters of their study is beneficial in the laser sintering process.

[0107] In stark contrast to these conclusions, the inventors identified extensive advantages to laser melting and laser sintering processes by employing low-porosity powders. These advantages are discussed below.

[0108] Reducing porosity increases the overall absorption of laser energy in powders, especially as the powder layer thins, for example, for layers less than 100 μm thick. This increased absorption means that more laser energy is used to melt the particles than is scattered in the print chamber environment, making printing with low-porosity powders more efficient. Since laser melting and laser sintering processes require expensive, high-power lasers, the ability to reduce the desired laser power can lead to significant cost reductions in laser components for printers. For example, in a comparison of the absorption rates of a standard titanium alloy powder with a skin thickness of approximately 65 μm and a porosity of approximately 40% and a low-porosity titanium alloy powder with a skin thickness of approximately 20 μm and a porosity of approximately 10% in a laser melting process, the low-porosity powder absorbs approximately 2.5 times more laser energy, and therefore allows the use of 60% lower laser power.

[0109] Reducing porosity also reduces the lateral scattering of laser energy during lateral heating of powders, resulting in more focused heating. This leads to a more integrated and controlled melting or sintering process.

[0110] Reducing porosity exponentially increases the thermal conductivity of a powder. This is partly due to the increased number of conduction pathways present in low-porosity powders, which have more contact between particles. The increase in thermal conductivity from a 40% porosity powder to a 5% porosity powder can be 10 to 50 times, and this change can result in numerous improvements for laser melting and laser sintering processes.

[0111] One improvement resulting from this increased thermal conductivity is a significant reduction in the temperature gradient present in the powder layer during heating, melting, and resolidification. This reduced gradient can be as large as 250,000 Kelvin / cm or more. A larger temperature gradient means greater thermal expansion in some parts of the liquid, and therefore greater volume change from liquid to solid, leading to more uneven resolidification (including cracks) at particle and inter-particle boundaries in metallic solids. Therefore, powders with increased thermal conductivity allow for less resolidification at particle boundaries and cracks. This leads to greater strength and greater durability of the solid under strained, high-temperature, and corrosive conditions, as well as improved surface structures of parts for more heat flow and therefore melting.

[0112] The reduced temperature gradient due to the high thermal conductivity of the powder also extends to the metal material beneath the powder. This means that more already printed layers will be heated to higher temperatures when the powder has relatively high thermal conductivity, resulting in more annealing of the metal, a reduction in particle boundaries, and generally a more integrated microstructure of the final part. This also leads to a reduction in the undesirable "bead formation" on the surface of the liquid metal, which can lead to pore formation in the final printed solid, as the interfacial energy decreases below a low temperature gradient across the entire interface.

[0113] The relatively high thermal conductivity of powders also leads to the relatively low maximum temperature exhibited during laser melting or laser sintering processes. The difference in maximum temperature can exceed 1,000 Kelvin, where the maximum temperature at the powder surface (liquefaction and sometimes boiling) occurs. Relatively high maximum temperatures imply stronger oxidation, which is undesirable as it leads to the introduction of oxidation-induced spherical and oxidative impurities. Relatively high maximum temperatures also imply more "splashing" of the surface liquid, resulting in a rougher surface and the possibility of micro or large pore formation. Pore formation leads to brittleness, reduced durability, and decreased corrosion resistance of solid components. Relatively high maximum temperatures further imply that in alloys, phase separation is more likely, for example, in liquid stainless steel, leading to the migration of carbon to its surface. Phase separation is undesirable as it leads to brittle structures and even delamination of layers.

[0114] Reducing porosity also reduces the amount of ultra-high temperature gas present in the voids between particles in the powder structure. This is important because the gas must be released during the melting and resolidification processes. Gas release during solidification can lead to the formation of micro and large pores, along with particle boundary formation.

[0115] Non-limitingly, for the advantages including those discussed above, and due to the inconclusive nature of previous studies examining the effects of employing relatively high-density powders in additive manufacturing processes involving powders, and because it has been impossible to date in general practice to produce practical powders (meaning sufficiently low viscosity and minimal phase separation) with a porosity of 20% or less, the inventors propose a general process improvement for additive manufacturing processes that involve sintering or melting, in the formation of a continuous final product in which layers of composite materials exhibit a porosity of 20% or less using powders. Non-limitingly, methods for producing such powders from a different range of materials, including metals, alloys, ceramics, cermets, and glass, in the process of increasing the density of granular composite materials are described above. Specific examples of methods for producing practical powders with a porosity of 20% or less using the process of increasing the density of granular composite materials are presented in the Examples section. Specific examples of methods for producing practical powders with a porosity of 20% or less using other processes can also be found in the Examples section.

[0116] When employing powdered metal inks corresponding to direct metal fusion sintering (DMLS) or selective laser fusion (SLM) processes, print speeds can be dramatically improved by using inks produced according to the density enhancement process for granular composite materials described herein. It is useful to compare the “performance-enhanced ink” with standard metal powders consisting of roughly the same size, nearly spherical particles (in the exemplary manner). Assuming that the melting time of this ink is a limiting factor in print speed, a minimum increase of 50 to 100 times in print speed can be expected due to the increased thermal conductivity and lower porosity of the performance-enhanced ink. These calculations assume a natural background gas of Ni or Ar at approximately 1 atmosphere, and for Ti, Ti alloys, stainless steel, copper, nickel-based superalloys, aluminum oxide, tungsten carbide, and 1 W / m². *It is effective for various metals, ceramics, carbides, and other inks, including other materials with a volumetric thermal conductivity greater than K. The increase in speed is more pronounced at lower pressures and for materials with higher volumetric thermal conductivity. Increases in print speed are accompanied by significant increases in volume, shear, and Young's modulus (i.e., increased mechanical strength), and significant increases in electrical and thermal conductivity.

[0117] When we compare the “performance-enhancing ink” to a metal powder consisting of a mixture of two sizes of roughly spherical particles (in the illustrative manner), the larger particles are considered to be approximately 100 times larger in volume than the smaller particles. Assuming that the melting time of this ink is a limiting factor in print speed, a 5 to 25-fold increase in print speed is expected due to the increased thermal conductivity and low porosity of the performance-enhancing ink. These calculations assume a natural background gas of N2 or Ar at approximately 1 atmosphere, and for Ti, Ti alloys, stainless steel, copper, and 1 W / m². * This is effective for various metallic inks, including any other metallic components or alloys with a volumetric thermal conductivity greater than K. The rate increase is more pronounced at lower pressures and for materials with higher volumetric thermal conductivity. The increase in print speed is accompanied by a significant increase in volume, shear, and Young's modulus (i.e., increased mechanical strength), and a significant increase in electrical thermal conductivity. Furthermore, due to the reduction in porosity and defects, a significant improvement in the reproducibility of high-quality printed products is expected in the final printed product. The increases in mechanical strength, electrical conductivity, and reproducibility are not expected to be as extreme compared to those seen with general powder inks.

[0118] In describing performance-enhancing inks, the most important point to remember is that even with the same porosity (porosity is interchangeable with density if a single alloy or component is identified), different microstructures can result in different characteristics, including viscosity, conductivity (related to print speed), and mechanical strength. [Examples]

[0119] experiment Example 1. This is an example of producing high-density powder in an additive manufacturing process that does not use the GCDEP in question. The material is Ti6-4 (Ti-6Al-4V, or titanium containing approximately 6% aluminum, 4% vanadium, and several trace components, including iron and oxygen, by weight). In this example, two groups of commercially pure Ti6-4 particles with significantly different average particle sizes are used to produce compressed metal powder with a porosity of approximately 18%. The first particle group exhibits a nearly normal distribution with an average particle volume of 8,000 μm³ (effective diameter of approximately 25 μm) and a small diameter standard deviation of 5 μm. The second group's volume follows a nearly log-normal distribution (meaning a nearly normal distribution passing curve as a log function of the effective particle diameter or particle volume, and also a volume probability density function) following the effective diameter, with an average particle volume of 0.5 μm³ (effective diameter of approximately 1 μm) and a diameter standard deviation of 0.5 μm. Both particle groups exhibit high sphericity (sphericity > 0.95), and the frictional interaction coefficient between particles within the same group is approximately 0.28, indicating the presence of a thin native oxide film (thickness < 5 nm) surrounding the particles. The incompressible packing ratio (percentage of space occupied by particles) for the first group is 0.60, and for the second group it is 0.595, which is consistent with the sphericity and frictional coefficient of those groups. When thoroughly mixed without vibration, the porosity of the volumetric material would be approximately 18%, such that the mass ratio of the larger group to the smaller group is 71% to 29%. For example, when vibrated, these particles tend to undergo phase separation, so the exact porosity depends heavily on the mixing method. A phase-separated mixture would exhibit a porosity of approximately 40%. To achieve a porosity of 18%, thorough up-and-down mixing with minimal vibration and a compression step are necessary. For example, in additive manufacturing, suppose a 100 μm thick layer of mixed powder is deposited using a roller, the height is slightly reduced by this roller, and then the mixture is compressed by applying a second roller to the deposited layer. In this case, the surface of the layer will show a higher porosity than in the agglomerated state, but the depth will be about 10 to 20 μm.

[0120] Example 2. This is an example of producing high-density powder for additive manufacturing without using the GCDEP in question. The material is 316 stainless steel (approximately 16.5% carbon, 12% chromium, 3% nickel, 1.4% molybdenum, 0.8% silicon, and trace amounts of phosphorus, sulfur, and other components). In this example, a single group of commercial-purity stainless steel particles exhibiting a continuous, nearly log-normal distribution size (diameter) probability density function with an average effective particle diameter of approximately 50 nm (effective volume of 65,000 nm³), a sphericity of 0.86, and a friction coefficient of 0.52 is used to produce powder with approximately 20% porosity. To achieve this, the geometric standard deviation of the particle group must be approximately 5.5 μm, meaning that 5% of the total particle volume is larger than 1 mm in diameter (and 40% are larger than 100 μm in diameter). In this case, since there is no concern about phase separation, a vibration step would be effective in reducing porosity for additive manufacturing. However, the laser process for a 1 mm layer is complex, and the amount of time required for the laser to heat the bottom of the layer 1 mm from the initially irradiated surface to its melting temperature hinders the uniform melting of particles (with minimal evaporation). Furthermore, smaller nanometer-scale particles have a very high specific surface area relative to their volume and are unrestrictedly exposed to the influence of forces including electrostatic, Casimir, and van der Waals forces, causing them to "adsorb" to each other and to larger particles. To neutralize electrostatic forces, a minimal residual charge environment ("electrostatic-free" environment) is crucial, and an additional compression step would likely be necessary to achieve a 20% porosity.

[0121] Example 3. This is an example of the production of a high-density granular composite material using the GCDEP. Two groups of lead glass (also called soda-lime glass) beads with high sphericity >0.98 and low static friction coefficient <0.05 were mixed to form a filter visible to the naked eye. The larger bead group had a diameter of 10 mm, and the smaller group had a diameter of 2 mm. The bead mixture was very uniform, i.e., the standard deviation of the volume probability density function for each group was almost zero. Table 1 is a list of numerous critical porosities corresponding to simulated mixtures of frictionless spheres (simulations were performed using the TJ algorithm) in ratios of the average diameter of small spheres to the average diameter (and volume) of large spheres. Referring to Table 1, we find that the minimum porosity of 21.6% appears at a relative volume fraction of 20.6% for small spheres. No compression or vibration steps are required to achieve this porosity, and excessive vibration would lead to phase separation of the beads at these size ratios. Upon mixing, a minimum porosity of 21.4% was found, appearing within 0.9% of the simulated value, and a relative volume fraction of 20.1% for the small spheres was found, appearing within 2.4% of the simulated value. The slightly lower porosity found in the relative volume fraction of the smaller spheres appeared to be a result of some hierarchy of the larger spheres appearing during mixing. Figure 7 includes an image showing the laboratory setup for these tests. The beaker in this image contains a mixture of 2mm beads with a relative volume fraction of approximately 17.5% and 10mm beads with a relative volume fraction of 82.5%, resulting in a porosity of 22.2%. This image does not capture the minimum porosity structure corresponding to this diameter ratio of beads.

[0122] For a mixture of 1mm and 10mm beads exhibiting the same physical characteristics as the aforementioned beads, a minimum porosity of 17.7% was found with a relative volume fraction of 25.0% small beads. This is within 0.6% of the minimum porosity of the small spheres predicted by the simulation, and is the smallest value to appear within 0.4% of the critical relative volume fraction of small spheres (data shown in Table 1).

[0123] For a mixture of 1mm and 3mm beads exhibiting the same physical characteristics as the aforementioned beads, a minimum porosity of 28.8% was found with a relative volume fraction of 24.5% of the spheres. This is within 0.7% of the minimum porosity of the spheres predicted by the simulation, and is the smallest value to appear within 0.4% of the relative volume fraction of the spheres (data shown in Table 1).

[0124] Example 4. This is an example of the production of a high-density granular composite material using the GCDEP. Two groups of tungsten disulfide particles with high sphericity (sphericity > 0.95) and very low friction (static friction coefficient equal to 0.07, kinetic friction coefficient equal to 0.03) were mixed to minimize porosity. The first group exhibited a normal diameter distribution with a standard deviation of 0.9 μm and an average size of approximately 10.1 μm, and the second group also exhibited a normal diameter distribution with a standard deviation of 0.14 μm and an average size of approximately 1.4 μm. Because their friction coefficients are very low, their sphericity is very close to 1, and their standard deviation as a percentage of the average particle size is very small, these particle groups can be considered as frictionless spheres. Referring to Table 1, it is determined that the mixture minimizing porosity appears with a porosity of 18.7% at a relative volume fraction of 22.4% small particles. These values ​​are determined by linear extrapolation from the values ​​given in the table to the diameter ratio of 1.4 / 10.1 = 0.137, corresponding to the ratio of small particle diameter to large particle diameter, 0.10 and 0.15. Due to the low coefficient of friction, vibration or compression is not necessary to achieve a low porosity; however, a compression step may sometimes be necessary to counteract any electrostatic forces that occur in the mixing procedure without inducing phase separation.

[0125] Example 5. This is an example of the production of a high-density granular composite material using GCDEP. Three groups of highly aspherical, high-friction-coefficient particles, consisting of crushed stone and sand, are mixed with a cement-water paste to create concrete. In this example, the amount of cement required to form the concrete is minimized, meaning the concrete remains "workable" and still fills the spaces around the reinforcing bars without leaving visible pores after flow and compression. The largest particle group has an average roundness of 0.21 and a density of 2.79 g / cm³. 3 The first particle group consists of very elongated crushed granite with a density of 0.56, an average sphericity of 0.56, and a static friction coefficient of 0.92. Its particles are sieved between 4.00 mm and 4.18 mm sieves, yielding a number-average size of 4.09 mm with a nearly uniform size distribution from 4.00 mm to 4.18 mm. The porosity (of its own) of this particle group is P1 = 52%. The second particle group consists of natural sand with an average roundness of 0.54, a density of 2.66 g / cm³, an average sphericity of 0.76, and a static friction coefficient of 0.70. Its particles are sieved between 1.00 mm and 1.25 mm sieves, yielding a number-average size of 1.13 mm with a nearly uniform size distribution from 1.00 mm to 1.25 mm. The third group, also natural sand, exhibits the same characteristics as the second group, except that it can be sieved with 0.15 mm and 0.20 mm sieves, producing a nearly uniform particle distribution from 0.15 mm to 0.20 mm with an average size of 0.18 mm. The second group has a porosity of P2=45%, and the third group has a porosity of P3=44%. The fourth particle group consists of Portland Gray cement, which has an average roundness of 0.78 and a density of 3.15 g / cm³. 3 It exhibits a density, an average sphericity of 0.88, and a static friction coefficient of 0.38. Its volume probability density function is approximately 97,500 μm. 3 An arithmetic standard deviation of approximately 33,500 μm (in units of linear size, 57 μm), which follows a nearly log-normal distribution. 3The average particle volume (effective linear size (diameter) is 40 μm) means that 97% of the cement particles have a linear size smaller than 150 μm (69% are smaller than 75 μm). The porosity of the cement is 30%. However, the size distribution and average particle volume of the cement change when the cement chemically reacts with water (the particles become smaller), and in this case, the size distribution and average particle volume are negligible. This is possible when considering the reaction with water, however, if the smallest sand particles are more similar in size to the cement particles, then the average particle volume and size distribution of the cement cannot be ignored.

[0126] These three particle groups are divided into subsets using the method shown in Figure 6, generating two subset particle groups, {1,2}1 and {2,3}2. In this case, the relative volume fraction of the small particle group at which the minimum porosity appears, corresponding to each subset, is experimentally determined using the experimental method described in subsection B above. As mentioned above, no vibration or compression steps were used to measure the volume fraction. For subset 1, the minimum porosity is 44% when the relative volume fraction of the second group of particles is 40%. For subset 2, the minimum porosity is 26% when the relative volume fraction of the third group of particles is 24%. Using the second step of the GCDEP described in subsection B above, it is found that the volume fraction of the first group of particles is φ1 = 0.56 * 0.60 = 0.336, and the volume fraction of the second group of particles is found to be φ2 = 0.56 * 0.40 = 0.224. Volume fraction V 00 =0.398, and V 01 = 0.266, and φ3 = V 00 * 2,3 +V 01 This yields (1-P3)=0.220. The volume fraction of the third group of particles is found to be (1-0.364-0.358)*0.56=0.156. The porosity of this aggregate mixture is 22.2%.

[0127] Cement paste is added to a mixture of three aggregate groups along with water. The amount of paste should be at least sufficient to fill the voids between the aggregates; however, generally, more paste is added to allow the water-containing concrete mixture to flow. Since the properties of the cement paste depend on the amount of water mixed in, cement and water are usually combined in a predetermined mass ratio. However, it is necessary to add excess water to account for the water absorption of the aggregates. Aggregate groups 1, 2, and 3 absorb 1.1%, 1.5%, and 1.6% of the mass of water, respectively. In this case, cement was added at a mass ratio of 11.9% of the total mixture mass, along with 0.45g of water per gram of cement and an additional 0.09g of water per gram of cement to account for aggregate absorption. Assuming that cement and water maintain their densities in mixture (i.e., ignoring chemical reactions), the mass ratio of cement needed to fill the voids in the aggregate (with water added in a mass ratio of 0.54:1) would be 10.0%. However, since excess cement is added to improve viscosity, the final mass ratios of the components were 36.1% vs. 23.0% vs. 22.5% vs. 11.9% vs. 6.5% for group 1 vs. group 2 vs. group 3 vs. cement vs. water. These ratios represent a total of 74.6% aggregate, 9.4% cement, and 16.0% water by volume.

[0128] Example 6. This is an example of the production of a high-density granular composite material using the GCDEP. Three particle groups—a first and second structural group consisting of spherical cast tungsten carbide (WC) powder, and a third binder group consisting of cobalt (Co) powder—are mixed to form a composite material powder corresponding to a selective laser sintering method. These WC particles have high sphericity (average sphericity > 0.97) and exhibit a low static friction coefficient of 0.08 with the addition of a small amount of lubricant, which burns at a low temperature (< 500°C) during sintering. The first particle group has an average particle volume of 114 μm and a uniform size distribution from 111 μm to 118 μm. The second particle group also has an average particle volume of 23 μm and a uniform size distribution from 20 μm to 26 μm. Considering the high sphericity and low friction coefficient in these first two particle groups, these groups can be estimated to be frictionless spheres. The cobalt particles are rounded, slightly aspherical particles with a sphericity of 0.84 and a static friction coefficient of 0.37. Their volumes follow a nearly log-normal distribution with an average diameter of approximately 0.4 μm³ (effective diameter of 0.9 μm) and an arithmetic standard deviation of 0.25 μm³ (effective diameter standard deviation of approximately 0.78 μm). These cobalt particles exhibit an incompressible porosity of 31%. Using the method described in Figure 6, these three groups are divided into subsets, producing two subsets {1,2}1 and {3}2, the third group being assumed to be small enough to fit into the space left by the two larger groups.

[0129] In this case, the thickness of the powder layer sintered at each step of the additive manufacturing process is between 250 μm and 350 μm. To minimize variations in the surface thickness of the layer, the minimum porosity is not required in the mixing of the first two sets of particle groups; rather, it is required that the mixture contain 35% (by volume) 23 μm particles, thereby diluting the larger particles and reducing variations in the outer surface thickness. Table 2 is a list of porosity values ​​for different relative volume fractions of small spheres, corresponding to a simulated mixing of frictionless spheres in ratios of various average diameters of small spheres to average diameters of large spheres. Referring to Table 2, it is determined that the incompressible porosity of the mixture of the 65% first group and the 35% second group is 24%. In an ideal processing process, the molten cobalt penetrates and completely fills the pore spaces in the WC mixture. For uncompressed cobalt powder with a 31% porosity, the volume ratio of group 1, group 2, and group 3 particles must be 0.65, 0.35, and 0.35. Expressed as mass ratios, this is 55%, 30%, and 15% of group 1, group 2, and group 3 particles. Since the cobalt particles have a relatively small average particle size compared to the group 2 WC particles, a compression step would likely be necessary before sintering to minimize phase separation of the cobalt particles from the WC particles.

[0130] It should be noted that in most processing processes, due to differences in the densities of liquid and solid cobalt, the material produced by laser sintering will be porous, and therefore will require less mass than the cobalt used in this embodiment.

[0131] Example 7. This is an example of the production of a high-density granular composite material using GCDEP. Four groups of Ti6-4 particles, each consisting of approximately 89.5% titanium, 6% aluminum, 4% vanadium, 0.3% iron, and 0.2% oxygen, and trace components, are mixed for powder formation compatible with the laser melting addition process. The first particle group exhibits an average sphericity of 0.97, a static friction coefficient of 0.25, and a diameter of 302 μm. 3240 μm with average particle volume (effective diameter of 8.3 μm) 3 From 380 μm 3 The particle volume (effective diameter from 7.7 μm to 9.0 μm) is distributed almost uniformly. The second group shows similar average sphericity and static friction coefficient, and 4.8 μm 3 With an average particle volume of 2.5 μm 3 From 7.1 μm 3 The particle volume (effective diameter from 1.7 μm to 2.4 μm) is distributed almost uniformly. The third particle group exhibits an average sphericity of 0.91, a static friction coefficient of 0.37, and 0.075 μm 3 (Effective diameter of 525 nm) Average particle volume and approximately 0.015 μm 3 The volume probability density function, with its standard deviation, follows a log-normal distribution. The fourth particle group exhibits a mean sphericity of 0.87, a static friction coefficient of 0.42, and a thickness of 0.000697 μm. 3 (Effective diameter of 110 nm) Average particle volume and approximately 0.000290 μm 3 The volume probability density function, with its standard deviation, follows a log-normal distribution. Using the method described in Figure 6, these four groups are divided into subsets, producing two subsets: {1,2,3}1 and {3,4}2.

[0132] With this known size distribution, sphericity, and static friction coefficient, simulations using the TJ algorithm generate the minimum porosity and relative volume function that appear for the subset. Since this composite material is compressed before processing, compression is also taken into account in the simulation, resulting in a first subset porosity of 12.3% for the first, second, and third groups with relative volume functions of 65.8%:19.7%:14.5%, and a second subset porosity of 23.4% for the fourth group with a relative volume function of 22.3%, where the particles of the fourth group themselves exhibit a porosity of P4 = 38.8%. Applying the second step of the GCDEP as described in subsection B above, V 00 =0.213 and V 01 This produces =0.0437, and with φ1=0.578, φ2=0.172, and φ3=0.127, φ4=V 00* 2,4 +V 01 This yields *(1-P4)=0.0589. This gives a final porosity of 6.4% when the relative volume functions of the particles in group 1, group 2, group 3, and group 4 are 61.7%, 18.4%, 13.6%, and 6.3%.

[0133] Example 8. This is an example of the production of a high-density granular composite material using GCDEP. Seven groups of high-purity (>99.0%) alumina powder (Al2O3) are mixed, compressed under high pressure (200 MPa), and then solid-state sintered at 1550°C for use as granular armor. This particle group has the following properties: 1. The first group consists of cylindrical shapes with a height of 12.2 mm and a diameter of 8.6 mm (sphericity of approximately 0.87), with a volume of 680 to 740 mm³. 3 (710mm 3 The particles exhibit a nearly uniform distribution (average volume). The coefficient of friction of these particles is 0.34. 2. The second group consists of materials with volumes ranging from 19.5 mm³ to 22.0 mm³. 3 (21.8mm 3 It consists of angular, aspherical particles (sphericity of 0.72) distributed almost uniformly within an average volume (effective average diameter of 3.4 mm). The coefficient of friction of these particles is 0.74. 3. The third group consists of materials with volumes ranging from 0.55 mm³ to 0.9 mm³. 3 It consists of angular, aspherical particles (sphericity of 0.75) distributed almost uniformly with an average volume of 0.73 mm³ and an effective average diameter of 1.1 mm. The coefficient of friction of these particles is 0.59. 4. The fourth group has an average volume of 0.0016 mm³ and 0.0004 mm³. 3 It consists of highly spherical particles (sphericity of 0.98) that are normally distributed with a volume of standard deviation (effective mean diameter of 145 μm). The coefficient of friction of these particles is 0.17. 5. The fifth group has a volume of 5100 μm³. 3 From 6050 μm 3 (5580μm 3It consists of highly spherical particles (sphericity of 0.97) distributed almost uniformly within an average volume (effective average diameter of 22.0 μm). The coefficient of friction of these particles is 0.21. 6. The sixth group is 5.2 μm 3 With a standard deviation of (effective average diameter of 5.1 μm), the average volume is 69 μm. 3 It consists of rounded, slightly aspherical particles (sphericity of 0.88) that follow a log-normal distribution. The coefficient of friction of these particles is 0.38. 7. The seventh group is 0.05 μm 3 With a standard deviation of (effective average diameter of 0.8 μm), the average volume is 0.27 μm. 3 It consists of slightly rounded, aspherical particles (sphericity of 0.84) that follow a log-normal distribution. The coefficient of friction of these particles is 0.47.

[0134] Using the method described in Figure 6, this group of seven particles is divided into subsets, producing five subsets: {1,2,3}1, {3,4}2, {4,5}3, {5,6}4, and {6,7}5. For each subset, the relative volume of the smaller spheres is chosen to be approximately 20% larger than the volume corresponding to the minimum porosity, in order to ensure that there are no large pores during compression. Simulations using the TJ algorithm are used to determine the desired relative volume of the first subset, producing a minimum compressible porosity of 14.5%. However, for the volume fraction adopted, the relative volumes of the first, second, and third groups are determined to be 46.4%, 25.9%, and 27.3%, respectively, resulting in a compressible porosity of 16.9%. For the second subset, experiments are conducted to determine the relative volume fraction and porosity curves for the fourth group of particles. The results show that the compressible porosity at which the relative volume fraction of the fourth group of particles appears at 28% is 18.8%, but for a subset of the mixture, the given porosity is 19.7% when fourth group particles with a relative volume fraction of 33.5% are used. For the third subset, under compression, the particles of both groups can be estimated to be frictionless spheres. Using lookup table 2, linear interpolation is used when the ratio of the small diameter to the large diameter is 0.15 to obtain a porosity of 19.9% ​​at which the relative volume fraction of the group 5 particles appears at 27%. For the fourth and fifth subsets, simulations using the TJ algorithm are incorporated, again using the relative volume fraction values ​​of smaller particles that are approximately 20% higher than the relative volume fraction value at which their minimum porosity appears. This approach produces porosities of 25.5% and 21.8% at which the relative volume fractions of even smaller particles appear at 29.5% and 25.2%, respectively. The compressible porosity P7 of particle group 7 was simulated to be 34.2%. Experimentally, a P7 of 33.9% was obtained and this value was used.

[0135] The subsets are combined using the six-step method described herein. The following values ​​are calculated. 1.V0=1, φ1=0.386, φ2=0.215, φ3=0.227 2.V 00 =0.425, V 01 = -0.0261. Following step 3b) of the subset combination technique, φ1 is reduced to 0.369 and φ2 to 0.206, V 01 We obtain =0. The value φ4 = 0.114. 3.V 000 =0.195, V 001 =0.0026, V 010 =0. Value φ5 = 0.0436 4.V 0000 =0.080, V 0001 =0.00070, V 0100 =0, V 0101 =0, V 0010 =0.0026. Value φ6 = 0.0480 5.V 00000 =0.030, V 00001 =0.0080, V 00010 =.00070, V 01000 =0, V 01001 =0, V 01010 =0, V 00100 =0.00099, V 00101 = 0.00026. Value φ7 = 0.0125. When the relative volume fractions from the largest particle group to the smallest particle group are 37.26%:20.75%:22.90%:11.54%:4.40%:1.88%:1.26%, it is found that the final porosity of the compressed powder is 0.90%. To achieve a porosity of 0.90%, careful mixing of the powder in an environment free from static electricity is required before compression. [Table 1] [Table 2]

Claims

1. Contains irregular, free-flowing powders that do not phase separate upon flow and have a void content of 25% or less. A layer of granular composite material, The thickness of said layer is less than 1000 microns. the law of nature, the granular composite material includes at least two groups of particles, each group being partitioned into mutually exclusive ranges of particle volume, the two or more groups comprising all of the particles in the granular composite material; the volume dividing the groups is the smallest minimum point in a pass curve or similar size distribution representing the granular composite material, and the value of the pass curve or similar size distribution at that minimum point is 75% or less of the value of the pass curve or similar size distribution at the largest maximum point of either of the two groups that sandwich the minimum point; the average particle volume of particles associated with the larger of said two groups is less than 10,000 times the average particle volume of particles associated with the smaller of said two groups; on average, particles of the larger group are in contact with at least one other particle of the same group, and the force network that maintains the mechanical stability of the granular composite is made up of a non-negligible amount of particles from both the smaller and larger groups, layer.

2. The powder is sinterable , fusible, or meltable powders A claim 1 The layer described in

3. The powder is metal particles , ceramic particles, cermet particles, or carbide particles Claims including 1 The layer described in

4. The powder is a mixture of ceramic particles and metal particles. , or a mixture of ceramic particles, metal particles and polymer particles Claims including 1 The layer described in

5. The powder contains titanium alloy particles, and the layer is 1 0% porosity. 1 The layer described in

6. The powder comprises first and second groups of particles, and the particles of the first group have a mean particle volume greater than or equal to the mean particle volume of the particles of the second group. Up to 2,000 times larger Average particle volume of claim 1 The layer described in

7. 10. The powder of claim 9, wherein the powder comprises first and second groups of particles, the particles of the first group having an average particle volume that is between 25 and 2,000 times greater than the average particle volume of the particles of the second group. 1 The layer described in

8. The powder comprises a first group containing 10 micron particles, a second group containing 2 micron particles, a third group containing 200 nanometer particles, and a fourth group containing 40 nanometer particles. , respectively. A mixture of 62.8% vs. 16.2% vs. 16.7% vs. 4.3% as The powder 4 0.4% porosity. 1 The layer described in

9. The layer is a powder Shape and a second layer of granular composite material, the second layer having a thickness of less than 1,000 microns, the granular composite material comprising: 25% Claims having a porosity of: 1 The layer described in

10. Both layers 、5 0 microns thick. 9 The layer described in

11. Contains an irregular, flowable suspension that does not phase separate upon flow and has a void volume of 35% or less. A layer of granular composite material, The thickness of said layer is less than 1000 microns. the law of nature, the granular composite material includes at least two groups of particles, each group being partitioned into mutually exclusive ranges of particle volume, the two or more groups comprising all of the particles in the granular composite material; the volume dividing the groups is the smallest minimum point in a pass curve or similar size distribution representing the granular composite material, and the value of the pass curve or similar size distribution at that minimum point is 75% or less of the value of the pass curve or similar size distribution at the largest maximum point of either of the two groups that sandwich the minimum point; the average particle volume of particles associated with the larger of said two groups is less than 10,000 times the average particle volume of particles associated with the smaller of said two groups; On average, particles of the larger group have at least one other particle of the same group as their nearest neighbor, and mixing of particles of the smaller group causes the smaller particle group to occupy the space given by a composite composed of only the larger particles, increasing the space occupancy of the mixture compared to a composition of only the larger particles. layer.

12. The aforementioned In suspension Powder can be sintered , fusible, or meltable powders A claim 11 The layer described in

13. The aforementioned suspension However, metal particles , ceramic particles, cermet particles, or carbide particles Claims including 11 The layer described in

14. The aforementioned suspension However, a mixture of ceramic particles and metal particles , or a mixture of ceramic particles, metal particles and polymer particles Claims including 11 The layer described in

15. The aforementioned suspension contains titanium alloy particles, is 1 0% porosity. 11 The layer described in

16. The aforementioned suspension comprises first and second groups of particles, and the particles of the first group have an average particle volume of: Up to 2,000 times larger Average particle volume of claim 11 The layer described in

17. The aforementioned suspension comprises first and second groups of particles, the first group of particles having an average particle volume that is between 25 and 2,000 times greater than the average particle volume of the second group of particles. 11 The layer described in

18. The aforementioned suspension The first group contained 10 micron particles, the second group contained 2 micron particles, the third group contained 200 nanometer particles, and the fourth group contained 40 nanometer particles. , respectively. A mixture of 62.8% vs. 16.2% vs. 16.7% vs. 4.3% as the suspension 4 0.4% porosity. 11 The layer described in

19. The layer , hang and a second layer of a granular composite material in a suspension state, the second layer having a thickness of less than 1,000 microns, the granular composite material comprising: 35% Claims having a porosity of: 11 The layer described in

20. Both layers 、5 0 microns thick. 19 The layer described in