METHODS FOR THE PRODUCTION OF LOW-DENSITY MICROSPHERES BACKGROUND OF THE INVENTION

Technical Field of the Invention

The present disclosure relates to the synthesis of microspheres. In particular, low-density microspheres are synthesized from mixtures comprising aluminosilicates in an entrained flow reactor under partially oxidizing conditions with a carefully controlled time-temperature profile. More particularly, tools are provided to calculate the conditions for producing microspheres having desired properties. Description of the Related Art Low-density microspheres produced from and/or collected from the ash resulting from coal combustion (i.e., cenospheres), have unique properties that can be utilized in many industries. For example, they exhibit high energy absorption, which results in protection against electromagnetic interference. They are used as fillers in composite materials, insulations, and paints, and to produce superior lightweight materials. In fact, the properties of cenospheres are such that available sources are currently insufficient to meet demand, and harvested cenospheres regularly command very high prices.

The source of most cenospheres at present is pulverized coal combustion for electricity production, from which they are merely a byproduct that is separated from the flyash. The chemical and physical properties of cenospheres, the feed properties and reaction conditions that control their formation, and the cenosphere formation mechanisms themselves are presently largely unknown.

Especially desirable would be a means of predicting the formation of cenospheres based on the feed (e.g., coal) composition and the reaction conditions, as well as evaluating other sources of cenosphere forming materials and the conditions affecting their formation. Unfortunately, in spite of extensive stud}', the factors that control cenosphere formation are not well understood at this time. In addition, knowledge of whether cenospheres are produced at coal-fired power plants and, if so, in what quantity, is rarely available and not predictable with any degree of certainty.

Accordingly, a substantial need presently exists for an increased supply of low-density microspheres, as well as further understanding of both the mechanisms of their formation and the engineering of practicable synthesis techniques. It is therefore an object of the present disclosure to describe methods for economical production of low-density microspheres, as well as methods for controlling and optimizing microsphere production to yield low-density microspheres with known and desirable properties, which may be predetermined.

BRIEF SUMMARY OF SOME OF THE PREFERRED EMBODIMENTS

It is thus an object of the present invention to overcome the deficiencies of the prior art and thereby to provide new and economical methods for the production of low-density microspheres using synthetic feed mixtures or feed mixtures derived from fossil fuel materials. In some embodiments, the fossil fuels may preferably include coals of all rank, waste coal streams, petroleum coke, bitumen, and other carbon-rich fuels. In other embodiments, the synthetic mixtures may include precisely formulated mixtures of clays, carbons, and binding agents. The process preferably comprises exposing the fossil fuel or synthetic mixture in an entrained flow reactor to a high temperature partially oxidizing environment for a carefully controlled residence time.

Some embodiments include the production of low density materials that can be used in a wide range of products, such as sound dampening materials comprising low-density microspheres for the auto industry and the building industry.

Some preferred embodiments provide methods for the synthesis of low-density materials with specific chemical and physical properties. Accordingly, certain embodiments relate to tailoring the properties of low-density microspheres manufactured from fossil fuels and synthetic mixtures to produce material with optimal properties for particular applications.

In accordance with certain embodiments, a method is provided for producing hollow spheres with a diameter of up to about 1000 μm from a primary glass-forming component comprising mineral matter derived from fossil fuels. In some preferred embodiments the primary component may preferably comprise aluminosilicate materials derived from coal. In such embodiments, the methods preferably comprise one or more of the following steps: selection of coal or coal waste material and separation of aluminosilicate materials; separation of mineral matter from the coal source to concentrate aluminosilicate clays; addition of optional materials such as transition metals for color enhancement; grinding and sizing the feed material to provide a consistent particle size to the process; heating the clay and coal mixture in an entrained flow combustor at the proper time-temperature history to promote the formation of hollow particles; rapidly quenching the particles so as to trap an optional blowing agent in the sphere create low- density microspheres. The quenching is preferably performed relatively rapidly and may preferably be accomplished using gas, water, a combination thereof, or other materials and techniques as are known to those skilled in the art.

Certain preferred embodiments provide methods for producing hollow spheres with a diameter of up to 1000 μm from synthetic precursor glass-forming mixtures. In such embodiments, the methods preferably comprise one or more of the following steps: selection of a primary component, which may preferably comprise an aluminosilicate; selection of a binding agent that

wets the primary component; selection of an optional blowing agent; selection of optional additive(s) to produce specific chemical and physical properties; grinding to preferably provide an even distribution of binder and bloλving agent throughout the particles; blending and/or slurrying the components in correct ratios; drying the formulation to preferably produce a hardened material; grinding and sieving the dried formulation to produce precursor particles preferably ranging from 25 to 250 μm in diameter; firing the particles preferably at a temperature sufficiently high to melt the aluminosilicate material and preferably to decompose or trigger the blowing agent; quenching the particles, preferably rapidly so as to trap the blowing agent in the spheres thereby creating a hollow particle. The primary component may preferably comprise any aluminosilicate material, including without limitation one or more of fly ash. Type C flyash, Type F flyash, bottom ash, slag, aluminosilicate clays, kaolinite clays, illite clays, bentonite clays, volcanic ash, ceramics, glasses, or any other type of aluminosilicate materials known in the art, providing that the conditions described by the methods disclosed in more detail below, are met satisfactorily. The most preferred primary component is chosen from the group consisting of flyash, Class C flyash, Class F flyash, illite clays, volcanic ash, glass, and combinations thereof.

The binding agent may preferably wet the primary component at high temperature and may comprise one or more of alkali metal compounds or other materials known in the art subject only to the desired functions of the binding agent as described in more detail below in accordance with the present disclosure. By way of illustration, the binding agent may comprise alkali silicates, such as sodium silicate.

Blowing agents are materials that can be incorporated into the precursor mixture and that lead to the development of cells through the release of a gas at the appropriate time during processing. The amount and types of blowing agents may influence the density of the finished product by its cell structure. The optional blowing agent may preferably comprise one or more materials known in the art which may volatilize or otherwise produce or release a blowing gas at conditions employed for the production of low-density microspheres in accordance with the present disclosure, as are known in the art.

In certain preferred embodiments, methods are disclosed for making low-density microspheres comprising firing a precursor comprising a primary component and a binder at a temperature such that the viscosity of the precursor is within a window where the calculated value of logio (viscosity, in poise) is preferably between 2.5 and 3.0, more preferably between 2.6 and 3.0, most preferably between 2.8 and 3.0 (inclusive), wherein the binder preferably has a calculated surface tension that is less than the calculated surface tension of the primary component at the firing temperature, wherein the primary component has a calculated surface tension at the firing temperature that is preferably within the range of 295 to 310 dyne/cm, more preferably within the

range of 300 to 305 dyne/cm (inclusive), and quenching the resulting low-density microspheres preferably within 200 msec, more preferably within 120 msec, and most preferably within 60 msec. Where an optional blowing agent is used, the blowing agent preferably is activated, by which is meant decomposes or reacts, at temperatures where the viscosity of the primary component is within the viscosity window. In certain preferred embodiments, the primary component, binder, optional blowing agent, and other optional components may have essentially any composition whatsoever, so long as the above-described criteria are met.

In certain preferred embodiments, methods are provided for producing hollow spheres with a diameter of up to 1000 μm from synthetic precursor particulate mixtures comprising silica or silica gel. In such embodiments, the methods preferably comprise selection of a fluxing agent, which may preferably comprise an alkali oxide; selection of a binding agent, such as has been described above; selection of an optional blowing agent, such as has been described above; selection of optional additive(s) to produce specific chemical and physical properties; blending and/or slurrying the above components in correct ratios; drying the formulation to preferably produce a hardened material; grinding and sizing the dried formulation to produce precursor particles preferably ranging from 25 to 250 μm in diameter; firing the particles preferably at a temperature sufficiently high to melt the fluxed silica or silica gel material and decompose or trigger the blowing agent; quenching the particles, and preferably rapidly so as to trap the blowing agent in the spheres thereby creating a hollow particle. In certain preferred embodiments, the fluxing agent and the binding agent may comprise the same material.

In certain preferred embodiments, the synthetic precursor particulate mixture may comprise silica gel and the fluxing agent/binding agent may comprise sodium silicate. The synthetic precursor may preferably be exposed to an oxidizing environment at a predetermined temperature and with a fluxing agent concentration calculated such that the viscosity of the precursor at the firing temperature is less than about 1000 poise. The surface tension of the binding agent may preferably be less than the surface tension of the silica gel/flux mixture at the firing temperature. The surface tension of the silica gel/flux mixture at the firing temperature may preferably be about 303 dyne/cm. The residence time of the precursor particles at the firing temperature may preferably be from 0.5 to 2.0 seconds. In some embodiments, the binding agent and the fluxing agent may preferably have about the same composition or preferably comprise the same compound. In certain preferred embodiments the binding agent and the fluxing agent may preferably comprise sodium silicate.

In certain preferred embodiments, methods are provided for determining the operating conditions for producing low-density microspheres. In some embodiments, methods comprise the steps of selecting a primary component, calculating the viscosity of the particles of the primary

component, preferably using the Kalmanovitch-modified Urbain model, at a predetermined temperature, calculating the amount of primary component particles having a calculated logio (viscosity, in poise) less than about 3, iteratively adjusting the predetermined temperature until a final temperature is obtained at which about 80 per cent of the primary component particles have a calculated logio (viscosity, in poise) less than about 3, and the surface tension of the primary component at the final temperature, preferably calculated using the Al-Otoom model, is about 300 dyne/cm, selecting a binder having a surface tension at the final temperature that is less than that of the primary component, and firing a precursor comprising the selected primary component and selected binder at the final temperature to produce low-density microspheres, and rapidly quenching the resulting low-density microspheres. In certain embodiments where an optional blowing agent is used, the blowing agent preferably is activated at about the final temperature.

As will be described in more detail below, the present invention thus provides several advantages over previously known techniques, including significantly lower priced low-density microspheres with the desirable qualities previously obtainable only by harvesting cenospheres from the byproducts of coal combustion.

BRIEF DESCRIPTION OF THE DRAVSTNG FIGURES

Fig. 1 is a block diagram illustrating a process for producing low-density microspheres from a fossil fuel such as coal.

Fig. 2 is a block diagram illustrating a process for producing low-density microspheres from synthetic materials.

Fig. 3 is a schematic diagram representing a down-fired tube furnace for production of low- density microspheres.

Fig. 4 is a schematic diagram representing the flow scheme of a down-fired tube furnace used for production of low-density microspheres. Fig. 5 is a cross-sectional diagram of a quench probe for use with a down-fired tube furnace for production of low-density microspheres.

Fig. 6 is a calculated viscosity histogram for a Texas lignite.

Fig. 7 is a plot of the surface tension of an aluminosilicate calculated as a function of temperature. Fig. 8 is a plot of calculated and measured viscosity of an aluminosilicate as a function of temperature.

Fig. 9 is a plot of calculated surface tension of an aluminosilicate ash and a binder as a function of temperature.

Fig. 10 is a combined plot of calculated surface tension for green illite and sodium silicate, and viscosity of green illite, all as a function of temperature.

Fig. 11 is a plot of calculated viscosity as a function of temperature for green illite. Fig. 12 is a scanning electron micrograph illustrating low-density microspheres. Fig. 13 is a plot of surface tension calculated as a function of temperature for silica gel in accordance with the present disclosure. Fig. 14 is a plot of viscosity calculated as a function of temperature.

Fig. 15 is a plot of bulk density as a function of drop height for low-density microspheres produced as described herein.

Fig. 16 is a scanning electron micrograph illustrating low-density microspheres. Fig. 17 is a scanning electron micrograph illustrating a low-density microsphere. Fig. 18 is a scanning electron micrograph illustrating a low-density microsphere.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Herein will be described in detail specific preferred embodiments of the present invention, with the understanding that the present disclosure is to be considered an exemplification of the principles of the invention, and is not intended to limit the invention to that illustrated and described herein. Thus the present invention should not be interpreted to be limited to the specifically expressed methods or compositions contained herein. In particular, various preferred embodiments of the present invention provide a number of different configurations and applications of the inventive method, compositions, and their uses. Suimnaiy of Known Techniques Cenospheres have a multitude of uses, including as fillers in composite materials, insulations, and paints, and in the production of superior lightweight materials. U.S. Geological Survey, Fact Sheet 076-01, Online Version 1.0, August 22, 2005.

Notwithstanding their desirable properties, the use of harvested cenospheres is limited by their relatively cost and low availability. Cenospheres occur in minute quantities. See, for example, U.S. Pat. Nos. 4,121,945 (less than 1% in flyash) and 4,268,320 (none at many plants); Glossωy of Terms Concerning the Management and Use of Coal Combustion Products

(CCPs), American Coal Ash Association, Inc., April, 2003 (2% maximum in flyash); Energeia,

University of Kentucky, Center for Applied Energy Research, Vol. 7, No. 4, 1996 ("minute concentrations", 0.1 1%). Harvesting cenospheres from fly ash is expensive, and many power plants simply do not produce significant quantities.

Several methods for producing or manufacturing microspheres are known. For example, U.S. Pat. No. 2,978,340 describes a method of forming glass microspheres from discrete, solid particles consisting essentially of an alkali metal silicate. The microspheres are formed by heating the alkali metal silicate at a temperature in the range of 1000-2500 ⁰ F in the presence of a gasifying agent, such as urea or Na ₂ COs. These alkali silicate microspheres

suffer from poor chemical durability due to a high percentage of alkali metal oxide. Early methods for manufacturing hollow glass microspheres suffered from the use of expensive starting materials, resulting in microspheres that were necessarily expensive. Poor chemical durability also limited their utility. Microspheres in various property ranges are also are generally available commercially from a variety of sources, for example 3M, Inc. (Scotchlite™ glass bubbles, Zeeospheres™), Akzo Nobel (EXPANCEL® microspheres), and PQ Corp. (EXTENDOSPHERES®). Most commercially available supplies continue to suffer from the deficiencies of high cost and low strength. Thus most applications that justify commercial manufacture relate to glass microspheres and high-cost applications. Such microspheres are unsuitable for use as, for example, high-strength construction fillers.

The prior art microspheres have numerous interesting combinations of properties but insufficient strength and economic viability to justify commercial interest. Manufactured glass microspheres are expensive with sub-optimal properties, while cenospheres harvested from fossil fuel-fired plants are (at best) merely expensive, if inconsistent. The approach of the present disclosure is therefore to economically produce low-density microspheres by preparing and processing precursor materials under carefully managed and predictable conditions. Cenosphere Composition

Coal combustion sources for flyash however, are generally not very consistent, and the quality varies significantly as a result. As reported by Raask, the microspheres formed in flyash are usually between 20 and 300 μm in diameter {Mineral Impurities in Coal Combustion, Hemisphere Publishing Co., Washington (1985)), though particles as large as 500 μm have been reported by Sokol et al. {Fuel Processing Technology 67, 35-52 (2000)). It is generally recognized that some method of bubble formation is necessary to produce porous microspheres. Shibaoka and Paulson {Fuel 65 (1986)) have identified the mechanisms for the formation, and Benson {Laboratoiγ Studies of Ash Deposit Formation During the Combustion of Western US Coals, Ph.D. dissertation, The Perm. State Univ., (1987)) suggested that CaCC> ₃ and carbon, when combined with clay minerals at high temperatures, could provide a viscous material and an agent to cause expansion.

The chemical characteristics of the microsphere precursor materials may thus be postulated to have an influence on how the microspheres form, as well as their subsequent chemical and physical properties. For example, the composition of the precursor should be such that it provides a viscous liquid phase when exposed to high temperatures. Further, the viscosity of the material should be within an optimum range that will allow for the expansion of the sphere as a result of the carbon oxidation or mineral decomposition. In addition, the chemical composition of the microsphere will influence the physical properties of the micropheres produced. The physical

properties include size, strength, shape, wall thickness, and color. Furthermore, additives such as transition metals may produce sites on the cenospheres surface that have catalytic activity, and may thus provide reactive surfaces for better applicability in catalyst products.

The inorganic elements in coal occur as discrete minerals, organically associated cations, and cations dissolved in pore water. The fraction of inorganic components that is organically associated varies with the coal rank. Lower-ranked subbituminous and lignitic coals have high levels of oxygen. These groups act as bonding sites for cations such as sodium, magnesium, calcium, potassium, strontium, and barium (other minor and trace elements may also be associated in the coal in this form). In higher-ranked coals, bituminous and anthracite, the inorganic components consist mainly of minerals.

Mineral grains are usually the most abundant inorganic component in coal. The major mineral groups found in coals, according to Raask, for example, include silicates, aluminosilicates, carbonates, sulfides, sulfates, and phosphates, as well as some oxides.

The behavior of the mineral grains associated with coal during combustion can be predicted only from detailed information on the abundance, size, and association of the mineral grains in the coal. In addition, the association of the mineral grain with the coal matrix must be determined and classified. A mineral associated with the organic part of a coal particle is said to be "included". A mineral that is not associated with organic material is referred to as an "excluded" mineral. The behavior of the organically associated elements, that is, those elements that are atomically dispersed in the coal matrix, must also be measured as to their abundance in the coal. The organically associated elements will react and interact with the other ash-forming constituents during combustion.

Methods to determine the inorganic composition of coal have evolved significantly over approximately the past 80 years. The early methods involved concentrating the inorganic components for chemical analysis using ashing or gravity separation techniques. These methods are fraught with errors and do not provide quantitative information on the association and abundance of inorganic components in coals. Zygarlicke et al. have recently reported more advanced methods such as computer-controlled scanning electron microscopy (CCSEM) (Scanning Electron Microsc. 4(3), 579-590 (1990)). Benson et al. describe chemical fractionation to quantitatively determine the abundance, size, and association of inorganic components in coals (bid. Eng. Chem. Prod. Res. Dev. 24 145 (1985)).

A technique that has shown much promise for quantitative determination of the mineral portion of the inorganic components in coal is SEM and microprobe (energy dispersive x-ray analysis) analysis. Over the past 15 years, this technique has been used much more rigorously to determine the mineral component in coal. In order to determine the

size, abundance, and association of mineral grains in both high-and low-rank coals, CCSEM and automated image analysis (AIA) are the preferred techniques used to analyze polished cross sections of coal epoxy plugs. The CCSEM technique is used to determine the size, shape, quantity, and quantitative composition of mineral grains in coals. Quantification of the type and abundance of organically associated inorganic elements in lower-ranked subbituminous and lignitic coals is currently performed by chemical fractionation. Chemical fractionation is used to quantitatively determine the modes of occurrence of the inorganic elements in coal, based on the extractability of the elements in solutions of water, 1-molar ammonium acetate, and 1 -molar hydrochloric acid. The filtered residues or solvent is analyzed after each leaching using x-ray fluorescence (XRF) to determine the percentage of each element remaining. The non-extractable elements are associated in the coal as silicates, aluminosilicates, sulfides, and insoluble oxides.

The inorganic coal components undergo complex chemical and physical transformations during combustion to produce intermediate ash species. The inorganic species comprise vapors, liquids, and solids. It is known in the art that the partitioning of the inorganic components during combustion to form ash intermediates depends upon the association and chemical-characteristics of the inorganic components, the physical characteristics of the coal particles, the physical characteristics of the coal minerals, and the reaction/combustion conditions. In order to predict the effects of inorganic constituents on combustion systems, much work has been undertaken in the prior art to elucidate the mechanisms by which the size and composition of intermediate ash species are formed.

The physical transformation of inorganic constituents depends on the inorganic composition of the coal and combustion conditions. As mentioned, the inorganic components can consist of organically associated cations, mineral grains that are included in coal particles, and excluded mineral grains. There is therefore a wide range of combinations of mineral-mineral, mineral-coal, mineral-cation-coal, and mineral-mineral-cation-coal associations in coal. These associations are unique to each coal sample.

The physical transformations involved in fly ash formation include 1) coalescence of individual mineral grains within a char particle, 2) shedding of the ash particles from the surface of the chars, 3) incomplete coalescence due to disintegration of the char, 4) convective transport of ash from the char surface during devolatilization, 5) fragmentation of the inorganic mineral particles, 6) formation of cenospheres, and 7) vaporization and subsequent condensation of the inorganic components upon gas cooling. It is known in the art that because of these interactions, the resulting ash typically has a bimodal size distribution. The submicron component is largely a result of the condensation of flame-volatilized

inorganic components. The mass mean diameter of the larger particles is approximately 12 to 15 μm, depending upon the coal and combustion conditions. Sarofϊm et al. refer to the larger-size particles as the residual ash because these ash particles resemble, to a limited degree, the original minerals in the coal (Combust. Sci. Technol. 16 187-204 (1977)). Processes such as ash mineral coalescence, partial coalescence, ash shedding, and char fragmentation during char combustion and mineral fragmentation all play an important role in the size and composition of the final fly ash. Loebden and others (J. Inst. Energy 119- 127 (1989)) and Zygarlicke and others (Prepr. Pap. — Am. Chem. Soc, Div. Fuel Chem. 35(3) 621-636 (1990)) indicate that three potential modes for fly ash generation can be used to describe fly ash particle-size and composition evolution. The first, "fine limit," assumes that each mineral grain forms a fly ash particle and that the organically associated elements form fly ash particles less than 2 μm. The second, "total coalescence," assumes one fly ash particle forms per coal particle. The third, "partial coalescence, suggests that the fly ash composition and particle size evolve because of partial coalescence. The coal particle size and mineral association have an influence on the size and characteristics of ash particles produced. Some coals such as lignites and subbituminous coals have exhibited a multi-modal size distribution with an intermediate size at 1-3 μm derived from the more refractory organically associated elements such as calcium. Benson et al. examined in detail the effect of coal particle size and association of mineral grains (Ash Formation and Deposition Smoot, L.D. (ed.) Amsterdam: Elsevier, Ch. 4 299-373 (1993)). Also see Helble et al. ("Transformations of Inorganic Coal Constituents in Combustion S3'stems" US Dept. of Energy Pittsburgh Energy Technology Center, PSI-1024/TR-1 141 (1991)).

The transformations of excluded minerals are dependent upon the physical characteristics of the mineral. Excluded minerals such as quartz (SiO ₂ ) can be carried through the combustion system with its angular structure still intact. Excluded clay minerals can fragment during dehydration, melt, and form cenospheres. The behavior of excluded pyrite depends upon its morphology. Some of the pyrite may be present as framboids. Framboidal pyrite may fragment more easily than massive pyrite particles. In addition, the decomposition of pyrite is very exothermic, and it transforms to pyrrhotite and oxidizes to FeO, Fe ₃ O ₄ , and Fe ₂ O ₃ during combustion.

Computer models such as ATRAN (for "Ash TRANsformations") have now been developed to predict the particle-size and composition distribution (PSCD) of ash produced during the combustion of coal. This technique uses advanced analytical characterization data, boiler parameters, and a detailed knowledge of the chemical and physical transformations of inorganic components during combustion to predict the PSCD of the

resulting ash. The PSCD of the ash directly impacts deposit growth, deposit strength development, and ash collectability.

Cenospheres are known to form from most coal burned in pulverized coal combustors. The amounts of cenospheres produced at power plants vary, and the assumed specific gravity of such cenospheres particles is less than 1.0. therefore the particles often float on water, such as ash ponds. Not all coal ash cenospheres float however, and this should be taken into consideration when considering cenosphere use.

It is generally recognized in the art that some method of bubble formation is necessary to produce cenospheres, but the exact mechanism is still unknown and several schemes may be involved. One set of mechanisms proposed in the literature requires the diffusion of carbon particles into a melted sphere of ash where either iron carbide, iron oxide, or silicon dioxide reacts with the included carbon to form carbon monoxide, according to Equations 1, 2, and 3: 2Fe ₃ C + SiO ₂ = Fe ₃ Si + 3Fe + 2CO Eq. 1

3FeO + 4C = Fe ₃ C + 3CO Eq. 2 SiO ₂ + 3C = SiC + 2CO Eq. 3

Other mechanisms have been suggested. See, e.g., Shibaoka and Paulson (Fuel 65 (1986)) suggesting that CaCO ₃ Or clay minerals could provide a source of gas (either CO ₂ or H ₂ O) for bubble formation.

There is also uncertainty in the prior work regarding important chemical relationships and their influence on cenosphere formation. For example, Raask had previously suggested that Fe ₂ O ₃ was important; however, Ghosal and Self state that the lighter color of cenospheres suggests that cenospheres contain lower-than-average amounts of Fe ₂ O ₃ , which is supported by chemical analyses (Fuel 74(4) 522-529 (1995)).

Although there are many unknowns involved in cenosphere formation, it is known that most pulverized coal combustion systems form some fraction of their ash as cenospheres. These small, hollow, thin- to thick-walled particles can either float or sink.

De Zeeuw et al. report that the thick-walled spheres have the advantage of strength, with a reported ability to withstand an exerted pressure in excess of 100,000 psi (about 690 MPa - the maximum hydrostatic pressure the testing machine being used could apply) (Proceedings of the American Coal Ash Association International Symposium Paper-76-

386, 386 (1976)).

The wide range of compositions found in ashes and slags derived from coal make slag viscosity predictions a challenge. However, many of the models available work well for specific compositions. Recently, seven viscosity models were evaluated for their ability to predict the viscosity for 20 coal ash slags. The models included three modified Urbain et

al. (Trans. J. Br. Ceram. Soc. 80 139-141 (1981)) models (Kalmanovitch et al. (Mineral Matter and Ash Deposition fi'om Coal, Bryers, R. W., and Vorres, K.S., Eds., Engineering Foundation: New York, 89-101 (1990)), Schobert (Lignites of North America, Elsevier: New York, 327-331 (1995)), and Senior et al. (Energy & Fuels 9, 277-283 (1995)) and four empirical regression models (Watt/Fereday (The Physical and Chemical Behaviour of the Mineral Matter in Coal under the Conditions Met in Combustion Plant; BCURA, Leatherhead, Surry, England, AWC (1969)), modified Watt/Fereday (Combustion Fossil Power Systems, Singer, J. G., Ed., Combustion Engineering: Windsor, CT, pp. C11-C16 (1981)), Hoy et al. (J. Inst. Gas Eng., 5 444-469 (1965)), and Sage and Mcllroy (J. Eng. Power 82 145-155 (I960)). The Urbain-derived models are based on the known behavior of network formers and modifiers, whereas the empirical models were formulated by fitting data for a few hundred measured viscosities. The Senior 50/50 model performed well when the oxidation state and content of iron placed limitations on other models. The Kalmanovitch and Sage and Mcllroy models appear most accurate for coals high in SiO ₂ and lower in Fe.

Low-density microspheres are currently directly (i.e., intentionally) produced for applications such as relatively high-cost, but low-strength "microbubbles" for limited uses such as pharmaceutical deliver}' mechanisms and as experimental nuclear fusion fuel targets. Currently, directly produced microspheres have relatively low strength and are relatively expensive, while naturally occurring byproduct cenospheres from coal combustion have high strength and are relatively expensive. Directly produced microspheres are made from high cost materials using high cost equipment; indirectly produced microspheres must be harvested from immense piles of flyash in which they exist in but minuscule fraction.

It is thus highly desirable to provide the properties of indirectly produced cenospheres at production yields more typical of directly produced microspheres. If successful, such a method would provide low-density microspheres having properties of broad utility with reasonable production costs, owing to the high production yield. It is axiomatic that the design of such high yield production methods benefits from further understanding of the mechanisms of low-density microsphere production. Nomenclature

The term "low-density microspheres" means roughly spherical particles, with diameter or longest dimension less than about 1000 μm, preferably less than about 500 μm, generally at least partially hollow, with a bulk density generally less than about 2.2 g/cm ³ . There are many types of low-density microspheres. The term "cenospheres" has historically been used to represent one type of material (hollow, glassy spheres) in fly ash that is thought to be produced

(heretofore unintentionally) by partial or complete melting of coal mineral matter during the combustion process. Such microspheres comprise at most 1-2 weight percent of fly ash, and perhaps much less. This is particularly the case if only cenospheres with a density that is less than 1 g/cm ^J ("floaters") are measured. Accordingly, the present study was undertaken with the aim of developing methods to increase the yield of low-density microspheres from various precursor materials, particularly from the byproduct flyash of coal combustion processes, which may already include some cenospheres, in economically advantageous processes.

The terms cenosphere and low-density microsphere are generally used interchangeably, although "cenospheres" often refers to low-density microspheres produced with flyash during pulverized coal combustion. Some types of low-density microspheres produced by commercial concerns are also referred to as microbubbles, or other similar terms often intended to suggest some proprietary properties. Formation of Glass Phases in Fly Ash

In certain preferred embodiments, the primary component preferably comprises a glass- forming material, or may preferably be selected based on its glass-forming characteristics. As is known in the art, silicates present in the fly ash of most Western low-rank coals are amorphous or glass-like. There are several models related to glass formation, but the one that is generally accepted by ceramists is the random network model. This model views glass as composed of three- dimensional networks or arrays that lack symmetry and periodicity. A material characterized by one of skill in the art as a good glass-forming material is preferably one in which the rate of crystallization is slow with respect to the cooling rate. This fact has led many investigators to try to correlate the nature of the chemical bonds and the geometric shape of the crystals to the ease of glass formation. However, a unified hypothesis of glass formation has yet to be developed. The rules adopted for the formation of oxide glass are based n the random network model: 1) the coordination number of the anion (oxygen) is low, and each oxygen is not linked to more than two highly charged cations such as Si ⁴⁺ ; 2) the coordination of the cation is generally low (3 or 4), and the network former (Si ⁴⁺ ) is small and highly charged, and the interaction of the cation and anion in glass causes the formation of very strong bonds; 3) the oxygen polyhedra share only corners and not edges or faces, to allow for flexibility; 4) in order to form a network, three of the four corners of an oxygen tetrahedral in the starting structure are shared. For example, by way of illustration and not limitation, quartz has all of the corners of its tetrahedral shared and is a very good glass former. Similarly, clay minerals have three out of four corners shared and are also good glass formers.

Glass-forming oxygen polyhedra are triangles and tetrahedral, and cations forming such a coordination are considered in the art as network formers. Alkali silicates form glasses easily, and the alkali atoms occupy random positions throughout the structure to provide charge neutrality.

The alkali and alkaline-earth elements added to a glass structure can be viewed as network modifiers which provide additional oxygen ions that modify the glass structure. This increases the oxygen-to-silicon ratio. The addition of sufficient alkali and alkaline-earth oxides can cause a breakup of the network, which eventually leads to crystallization. The conditions for glass formation are known in the art. Glass-forming oxides, sometimes referred to as network formers, have the ability to build and form three-dimensional, random networks. Oxides that are good network formers include, but are not lilmited to, SiO ₂ , B ₂ O3, GeO ₂ , P ₂ O ₅ , and AsO ₄ . These network formers form highly covalent bonds with the oxygen atoms. A network-modifying oxide is incapable of building a continuous network. The addition of such a modifying oxide to a network causes weakening of the network. Good examples of network modifiers are oxides of sodium, magnesium, calcium, and potassium. Addition of a network modifier to a continuous network it breaks up the network by adding oxygen to produce nonbridging oxygens. For example, the addition OfNa ₂ O to a silicate glass causes the formation of two nonbridging oxygens, one of which was contributed by Na ₂ O, and the sodium ion balances the charge. The addition of a network modifier to a silicate glass reduces the viscosity and increases the thermal expansion coefficient. Some intermediate oxides are not capable of forming a glass, but can take part in the glass network. A good example of an intermediate oxide is alumina. The addition of a network modifier results in the formation of nonbridging oxygens and weakening the network. By understanding the influence of various elemental species on the structure of glass, insight can be gained on the effects of various elemental components on the viscosity and surface tension characteristics of the liquid phases. The balance of network modifiers (such as alkali and alkaline earth elements) to network formers (such as aluminosilicates) affects the formation of low- density microspheres. Mechanisms of Formation Crystalline quartz and mullite along with a large amorphous glass phase are present in cenospheres and cenosphere-rich fly ashes. The chemical composition of a fluid slag able to produce cenospheres also seems to be a composition that permits the crystallization of mullite and quartz from the amorphous phase or viscous liquid phase. The quartz and mullite may form small widely dispersed macrocrystalline patches on the cenosphere surface as a result of the devitrification of the cooling amorphous liquid phase. Note that the types of crystals that form are dictated by the chemistry of the melt. The formation of quartz and mullite is characteristic of cenospheres derived from illite minerals, typically produced in pulverized coal combustion.

Cenospheres form commonly from both high- and low-rank coals and have been observed worldwide in thermal power systems. However, they seem to form within a relatively narrow

chemical composition range, and a composition expected to have viscosities that are quite high. The bulk fly ash has a similar chemical composition and viscosity to the cenospheres when cenosphere formation is significant. It is thus expected that the coal chemical composition will be similar to that of the cenospheres and fly ash for coals which produce significant numbers of cenospheres. It should thus, in theory at least, be possible to identify coals with the potential to produce cenospheres. Further, the narrow cenosphere composition range indicates that specific coal minerals are involved in cenosphere formation. These minerals are preferred for use in accordance with the present disclosure for economical production of low-density microspheres. Carbon dioxide, nitrogen, and water have been identified as being encapsulated in cenospheres, and reactions producing these gases suggested as mechanisms for cenosphere formation.

Without wishing to be bound by theory, we conclude that cenosphere formation depends on the following criteria: the liquid (e.g., molten ash) material preferably has a surface tension that is within a range sufficient to allow the formation of a molten particle and particle-to-particle wetting; a gas (blowing gas) is present or generated inside a molten ash droplet preferably in sufficient quantity to inflate the ash particle into a hollow cenosphere; the viscosity of the ash droplet is preferably low enough to permit the inflation to take place yet high enough so that the hollow droplet does not burst; the time-temperature history of the molten droplet is preferably such that the droplet is cooled before bursting and the cooling is preferably rapid enough that the hollow droplet does not collapse as the internal gas pressure decreases. Surprisingly, it has now been found that these criteria may preferably be calculated and predetermined according to some methods described in the present disclosure.

An important clue to the mechanism of cenosphere formation is the presence of gases trapped inside cenospheres. These gases are presumably the blowing agent which inflates a molten ash particle to form a cenosphere. From the literature on naturally occurring and synthetic cenospheres, CO ₂ , CO, N ₂ , H ₂ O, and SO ₂ are possible blowing agents. Volatilization of alkali inorganic species might also be considered a possible blowing agent. However, the range of cenosphere chemical compositions contains very little sodium, and the potassium appears to be incorporated in forms that are not volatile.

The presence of carbon dioxide may be the result of carbon oxidation, carbon gasification (along with CO), decomposition of carbonate minerals such as CaCO ₃ , or diffusion of CO ₂ into the molten slag. Carbon monoxide can result from carbon oxidation or gasification as well as mechanisms described subsequently. Water also can be the result of carbon gasification or oxidation, decomposition of hydrated and hydroxyl mineral species or outgassing of water trapped in pores of the coal. Sulfur dioxide could result from the decomposition of sulfate minerals such as CaSO4 or from the oxidation of sulfides such as FeS?.

A potential complication in elucidating the source of the blowing agent is subsequent reactions of the encapsulated gas. Water would be expected to react with alkali and alkaline oxides if present in the inner walls of the cenospheres to form the corresponding solid hydroxides such as CaO +H ₂ O -> Ca(OH) ₂ . Because flotation in water is commonly used to separate cenospheres, the presence of water could also be the result of infiltration through micro "pinholes" in the cenosphere walls, and not conclusively indicate water is the blowing agent. Sulfur dioxide may also react readily with alkali and alkaline earth oxides on the inner cenosphere wall to produce solid sulfate species, removing SO ₂ from the gas phase.

Equilibrium thermodynamic calculations were performed using the Facility for the Analysis of Chemical Thermodynamics (FACT) program, which is well known in the art and publicly available. FACT was used to calculate the equilibrium concentration of approximately

700 gas, liquid, and solid species as a function of temperature from 500° to 1500 ⁰ C in 25°C increments. The chemical composition used was based on the ultimate and bulk ash chemistry analyses of a typical bituminous coal. No iron carbide species were predicted to be thermodynamically favored over this temperature range. This suggests that an iron carbide mechanism is not likely to be involved in cenosphere formation.

Carbon dioxide can also be produced as the result of carbonate mineral decomposition, such as for the case of calcium carbonate:

CaCO ₃ → CaO + CO ₂ Eq. 4 In this case, there would be no CO produced.

Water is known to be the blowing agent for the production of some synthetic cenospheres ("shirasuballoons") from Shirasu volcanic glass. The formation process involves milling the volcanic glass and treatment with hydrochloric acid to remove surface material, presumably to assist the adsorption of water. Water is also used as the blowing agent in the production of some hollow ceramic oxide microspheres. Sulfur dioxide resulting from the oxidation of pyrite has been seen to produce hollow spherical particles similar to cenospheres, with the probable reaction being:

2FeS ₂ + 5.5O ₂ - Fe ₂ O ₃ + 4SO ₂ Eq. 5

Sulfur trioxide resulting from the decomposition of sulfate species is also used extensively as the blowing agent in the production of synthetic glass microspheres. Sodium sulfate is introduced into the molten high silica glass along with other additives to control viscosity. Control of the oxygen and SO ₃ partial pressure over the melt permits the manufacture of synthetic hollow microspheres with a wide variety of densities.

The presence of significant nitrogen gas encapsulated in cenospheres is difficult to explain. The reaction:

0.5Si ₃ N ₄ + 3Fe ₂ O ₃ -* 1.5SiO ₂ + 6FeO + N ₂ Eq. 6 has been suggested based on the bloating of feldspars due to the decomposition of small amounts of silicon nitride present. Other possible mechanisms to encapsulate N ₂ in cenospheres would be diffusion of N ₂ from the combustion gas as previously described or simple trapping of combustion gas as the molten ash droplet formed.

A possible mechanism not explicitly considered in the art is a combustion or gasification reaction involving char that is entrapped within a molten ash particle. Evidence exists that some bituminous coals form porous chars with burning within the particle and with fused ash particles present on the particle surface. Coal contains appreciable oxygen and hydrogen as well as carbon, as illustrated by the ultimate analysis of a typical Kentucky No. 9 coal, resulting in an empirical formula of C _{O O} HO _S O _{O I} NQ _OI S _{O CC} - Oxidation resulting from oxygen trapped as the slag becomes molten or oxygen diffusing into the slag would produce CO ₂ and H ₂ O along with small amounts of nitrogen and sulfur oxides. In an oxygen- deficient environment, gasification could also occur, producing CO and H ₂ as well as CO ₂ and H ₂ O. A typical product gas composition under strongly reducing conditions (in an oxygen-blown gasifier with no N ₂ present) comprises approximately 60% CO, 30% H ₂ , 8% CO ₂ , and 2% H ₂ O. Less stringent mildly-reducing conditions, as encountered in the slag layer of a cyclone barrel in a cyclone-fired thermal power plant, has an expected composition of about 75% N ₂ , 10% H ₂ O, 9% CO ₂ , and 5% CO, with small amounts of NO _x , SO ₂ and H ₂ S. The presence of CO, CO ₂ and especially H ₂ encapsulated in cenospheres would suggest that a gasification reaction is responsible for the blowing agent. However, it should be noted that entrapped oxygen cannot provide the blowing agent since one mole of CO ₂ is produced for each mole of O ₂ consumed, so there is no net change in gas volume. Modeling the Formation of Low-Density Microspheres The prediction of low-density microsphere/cenosphere formation holds the promise of reductions in production cost and improvements in product properties. Cenosphere formation is governed by the amount and rate of evolution of the blowing agent, the slag viscosity and surface tension, and the amount of slag material present (controlling ultimate size and wall thickness). Expressions for calculating cenosphere formation time are known in the art. The derivation is based on the rate of change in radius of a hollow sphere: dr/dt =r ² δP/4vs Eq. 7 where dr/dt is the rate of change in radius of a hollow sphere r is the radius of the sphere δP is the difference between the internal pressure P and external atmospheric pressure P _a v is the viscosity of the slag

s is the wall thickness

The value of δP is given by:

δP=P - Pa - 2γ (l/H- l/n) Eq. 8 where γ is the surface tension of the slag r is the external sphere radius ri is the internal sphere radius

The sphere is assumed to contain a certain amount of gas so that: P = k/ _{r ]} ³ Eq. 9

The value of k is determined from the previous equation by calculating a series of particle radius r with δP = 0 or some higher-pressure value assumed to occur when expansion has ceased.

The equations are combined into a single expression for the rate of sphere expansion: dr/dt = (l/(4v (r - r,))) ((kr7n ³ ) - P ₃ F ² ^ y (r/n)(r + >-,)) Eq. 10 Since the volume of slag present remains the same as the hollow sphere forms, r and rj are related by the expression: r ³ - η ³ = (3/4π)(slag volume) Eq. 11

The equation for dr/dt can then be used to calculate the rate of change at a given value of r and the formation time for a cenosphere. Certain assumptions must be made to set conditions for the calculation, namely: • The initial volume of slag present

• The initial volume and pressure of the gas inside (or the volume of evolved gas)

• The final pressure of the gas inside

• The time-temperature history undergone by the cenosphere as it forms

• The viscosity and surface tension of the slag which are governed by the temperature and slag composition

For the model of cenosphere formation described herein, the initial slag volume and composition were provided from the ash particle-size composition distribution calculated by the ATRAN program. The time-temperature history of the cenosphere formation is an input parameter dependent on the combustion/reaction system being modeled. The ash compositions may be used to calculate viscosities as a function of temperature, preferably using the modified Urbain equation (Kalmanovitch et al, Mineral Matter and Ash Deposition fi'om Coal, Bryers, R.W., and Vorres, K. S., Eds., Engineering Foundation: New York, 89-101 (1990)), which is known in the art.

The method of the Kalmanovitch-modified Urbain model for viscosity is briefly described below.

First, normalize the ash composition such that:

Na ₂ O + MgO + Al ₂ O ₃ + SiO ₂ + K ₂ O + CaO + TiO ₂ + Fe ₂ O ₃ = 100 Eq. 12 where the ash oxide components are expressed as mole fractions.

Then convert all the iron in the ash to equivalent FeO, and calculate M and α, where:

M = CaO + MgO + Na ₂ O + K ₂ O + FeO + 2TiO ₂ Eq. 13 α = M/(M + Al ₂ O ₃ + FeO) Eq. 14

Calculate B: B = B ₀ + (B ₁ SiO ₂ ) + (B ₂ (SiO ₂ ) ² ) + (B ₃ (SiO ₂ ) ³ ) Eq. 15 where:

B ₀ = 13.8 + 39.9355α - 44.049a ² Eq. 16

B ₁ = 30.481 - 1 17.1505a + 129.9978a ² Eq. 17

B ₂ = -40.9429 + 234.0486a - 300.04a ² Eq. 18 B ₃ = 60.7691 - 153.9276a + 211.1616a ² Eq. 19

Then calculate:

In(A) = -(0.2812B + 11.8279) Eq. 20

Finally: ln(μ) = In(A) + In(T) + 10 ³ B/T Eq. 21 where T is in °K and μ is in poise.

The surface tension may preferably be calculated as a linear function of the slag composition, using the method of Al-Otoom et al. {Energy & Fuels, (2000), 14, 994-1001): y r ∑ xi T i Eq. 22 where γ _t , is the total surface tension of the slag γ , is the surface tension of the i* individual component x , is the mole fraction of the i* component.

For some typical oxide components in cenospheres at 1350 ⁰ C: SiO ₂ has a surface tension of about 303 dyne/cm, Al ₂ O ₃ about 313 dyne/cm, CaO about 310 dyne/cm, Fe ₂ O ₃ about 311 dyne/cm, MgO about 314 dyne/cm, K ₂ O about 286 dyne/cm, and Na ₂ O about 298 dyne/cm, as given in the Al-Otoom et al. (2000) publication cited above.

The surface tension has a weak dependence on temperature, so it can be approximated by a linear function of temperature using experimental measurements at a specific

temperature as a reference point. The surface tension is estimated to vary by about 4 dyne/cm for every 100 ⁰ C change in temperature.

The volume and pressure of the gas in the forming cenosphere represent the major unknown at this point. The final pressure in the cenosphere is estimated from the expression: δP = 4 γ / r Eq. 23 where the variables are as previously defined. The excess pressure for the cenospheres is in the range of 0.09-0.58 atm (about 10 to about 60 kPa), with these values referring to room- temperature conditions. A refined estimate of the final internal pressure may be obtained form actual cenosphere size and composition data. Initial pressure and gas volume may then be estimated by back calculation. Alternatively, either the initial pressure and gas volume, or the rate of gas evolution, can be used as input variables, and a variety of gas evolution scenarios calculated. Low-Density Microspheres - Methods of Production

Low-density microspheres may be produced from a wide range of materials. The most common type of microspheres preferably comprise aluminosilicate glass materials. The aluminosilicate glass materials may preferably be derived from mixed layered clays such as illite, montmorillonite, and kaolinite, although other sources as are known in the art may also be used successfully. These materials may preferably rely on the presence of other elements such as sodium, potassium, calcium, magnesium, and iron as fluxing agents to reduce the viscosity of the materials and allow them to flow. These minerals are abundant in fossil fuels, and particularly in coal. In addition, various transition metals may preferably be added to these minerals to change the chemical and physical properties of the low-density microspheres.

Referring now to Figure 1 , there is shown a flow diagram of an embodiment to concentrate microsphere precursors, add transition metals, and produce low-density microspheres. Fig. 1 shows mineral separation in block 10 that may preferably be conducted with an air jig or similar device to separate the major minerals. A density separation process for clay minerals may preferably be incorporated into the air jig, as shown in Block 12. The density separation process may preferably be adjusted to provide a concentrated stream of clay-rich fraction. Block 14 preferably involves the addition of transition metals to modify the chemical and physical properties of the clay materials, which are preferably added to the clay-rich fraction at Block 16. Block 16 preferably comprises size fractionating the combined material to provide feed streams of specific particle sizes to allow for the optimum formation of microspheres. Block 18 describes the preparation of low-density microspheres by reacting the product of block 16, preferably in a high temperature entrained flow reactor. The high temperature entrained flow reactor may preferably be adjusted to the composition of the precursors to provide the optimum temperature and time to produce the desired microspheres.

Block 20 describes represents separation of the low-density microspheres into preferred density (or other) fractions.

Referring now to Figure 2, a flow diagram is provided for low-density microsphere production according to another embodiment. The materials for the formulation are first chosen as indicated at block 21, preferably dependent on the desired application for the low-density microspheres to be produced. For example, the formulation may preferably comprise one or more aluminosilicate materials, one or more binding agents, and an optional blowing agent. Other general formulations may of course be employed without departing from the scope of the invention. The materials are then preferably mixed and ground to a specified size and slurried as indicated in Block 22. Block 23 then preferably involves drying the slurry to removal all volatile matter that could rupture the microspheres during heating. The dried material is then preferably ground and sieved to the appropriate size as shown in Block 24. The sized particles are then fired as indicated in Block 25, preferably in an entrained flow reactor that can be adjusted to control the residence time and temperature to produce low-density microspheres of desired properties. Rapid quenching of the microspheres is then preferably provided as shown in Block 26 to ensure that the particles do not burst. The microspheres then may preferably be separated according to size, density, color, and other properties, as indicated in Block 27.

With fossil fuel-based precursor materials, some preferred embodiments of the present disclosure for production of low-density microspheres may comprise the following method. The selection of coal or coal waste material is first made, followed by the separation of preferred aluminosilicate materials. Mineral matter is then separated from the coal source to concentrate the preferred clays. If desired, optional materials such as transition metals may preferably then be added for the enhancement of color and and/or other microsphere properties. The resulting material is then ground or comminuted and sized, preferably to provide a consistent and uniform particle- size feed material. This feed material is then heated in an entrained flow combustor preferably having a predetermined time-temperature history to maximize the formation of low-density microspheres. The low-density microspheres produced are then preferably rapidly quenched so as to trap the blowing agent in the solidifying microspheres, producing a dimensionally stable product. Quenching to 900°C is preferably accomplished within 200 msec, more preferably within 120 msec, and most preferably within 60 msec. Precursor Preparation

With respect to selection and preparation of precursor materials, some preferred embodiments for production of low-density microspheres comprise the following method. The identification and selection of the binding agent (and optional blowing agent, if used) is made preferably to ensure proper formation of the hollow microspheres. The components are preferably

careflilly selected to prevent bursting of the forming microspheres, and also preferably to prevent the release of the blowing agent before the particles that will form the product microspheres are in the proper viscosity range for formation. By way of illustration and not limitation, binding agents may include Class C fly ash, sodium silicate, sodium sulfate, sodium borate, boric acid, and others. Illustrative blowing agents may include carbonaceous materials or any other materials capable of volatilization at appropriate conditions, as are known in the art. For example, the use of blowing agents in the preparation of foams, composite materials, and hollow balls is known in the art, as described in, e.g., U.S. Pat. Nos. 5,866,641 and 6,828,026. Suitable primary components may include aluminosilicate materials, Class C fly ash, Class F fly ash, illite and other clays, volcanic ash, glasses, silica gel, and others.

Selection of any optional additives to produce specific desired chemical and physical properties may then be made. This may preferably be based on the need for the specific application, such as the addition of a transition metal colorant. In some preferred embodiments, transition metals may preferably by added to change the color of the glass microspheres. In such embodiments, addition of the metals at very low levels, preferably less than 1 wt-%, will allow for modification of the microsphere color. For example, the addition of Mn, Cu, Fe, Co, Au, and combinations thereof may preferably be used to change the color of the microsphere. This addition may preferably enhance the use of the microspheres in specific applications.

In addition, any optional opalization agent could also be added. By way of illustration and not limitation, if it is desired to vary the light absorbing characteristics, iron or opalization agents such as SnO ₂ may also be added to create colored cenospheres.

Once the precursor materials have been optimized, the material is preferably ground or otherwise comminuted to provide a uniform and even distribution of binder and blowing agent throughout the particles. The aluminosilicate material may preferably be reduced to 5 μm to 8 μm diameter, and the blowing agent may preferably be reduced to less than 1 μm in diameter. In certain preferred embodiments where silica gel may preferably be used rather than an aluminosilicate, the above step may preferably be omitted. The selection of the type of blowing agent preferably depends upon the temperature at which the viscosity of the primary component is such that 2.8 < logio(μ) ≤ 3.0 (μ in poise). The blowing agent may preferably be selected so as to devolatilize and release blowing gas when the primary component is within the above viscosity range, and one skilled in the art will recognize that many suitable blowing agents may be used without departing from the scope of the invention. If the binder is an insoluble solid it may preferably be reduced to less than 1 μm in diameter.

The components may preferably then be blended and/or slurried in the desired ratio for microsphere formation. Poorly formed spheres may result from too much or too little of either the

07

binder or blowing agent. Preferably, the binding agent ranges from about 5 to about 25 wt-% (inclusive, dry basis), more preferably from about 5 to about 8 wt-% (inclusive, dry basis) of the total formulation. The optional blowing agent(s) preferably range(s) from about 0.1 to about 1% (inclusive, dry basis) of the total formulation, if used. The above formulation is then preferably dried to produce a hardened material. The dried formulation may then preferably be ground or otherwise comminuted, and sieved or otherwise size- classified, to produce precursor particles preferably ranging in size from about 25 to about 215 μm in diameter. The size of the precursor particles may preferably be varied to produce larger or smaller hollow microspheres. Narrower precursor particle size distributions are generally preferred over broader size distributions.

Hish Temperature Treatment

In some preferred embodiments, an entrained flow reactor that is co-fired with natural gas may be used to produce the low-density microspheres. The operating conditions, including firing rate, maximum temperature, gas residence time, particle residence time, and gas cooling rates are all preferably carefully controlled based on the composition of the precursor materials. The composition of the precursor material may preferably change depending upon the coal or waste coal source, or the aluminosilicate source.

The precursor particles are preferably fired at a temperature sufficiently high so as to melt the aluminosilicate material or silica material or other primary component material, and decompose (trigger) the blowing agent. The time-temperature history of the firing step is preferably carefully chosen to achieve the proper viscosity and surface tension conditions. The range of temperature for typical aluminosilicates, for example, is preferably from about HOO ⁰ C to about 1400 ⁰ C, and varies depending upon the properties of the precursor materials. The range of residence time is preferably from about 0.1 seconds to about 10 seconds, more preferably from about 0.2 seconds to about 2 seconds, and also varies with the properties of the precursor materials.

Referring now to Fig. 3, there is shown an atmospheric drop-tube furnace (ADTF) suitable for the production of low-density microspheres in accordance with a method of the present disclosure. The ADTF is a laboratory-scale, electrically heated, entrain ed-flow (down-flow), tube furnace with the ability to treat precursors, combust coal, and produce ash and/or low-density microspheres under closely controlled conditions. An important characteristic of the ADTF is the ability to accurately simulate the thermal history of (ash) particles in larger scale processes. A large variety of full-scale production reactor profiles can be quickly simulated to optimize full-scale conditions efficiently. The results obtained from ADTF testing are, therefore, expected to be applicable to larger scale entrained-flow furnaces for production of low-density microspheres. Preferably, entrained-flow furnaces for use with the methods of the present disclosure are operated

in a down-fired configuration, which is expected to provide the advantage of avoiding ash deposition in the vicinity of the burner(s), which can result in operational difficulties.

Accordingly, referring still to Fig. 3, the ADTF 100 is capable of maintaining gas temperatures up to 1600 ⁰ C (2912°F) in each of four zones, comprising preheat section 110, top furnace section 120, middle furnace section 130, and bottom furnace section 140. Precursor particles are injected to the furnace tube 101 via water-cooled inlet 111 entrained in a primary air or gas stream. Secondary air or gas is provided to the annular space 121 and flows through flow straightener 115 to furnace tube 101. Combustion/reaction parameters — such as initial hot-zone temperature, excess air, residence time, and gas-cooling rate — can be closely controlled and monitored. Treated materials, chars, ash, or microspheres can be collected after exposure to temperatures for various residence times to precisely control the time-temperature history.

Fig. 4 provides an overall schematic of the ADTF system, comprising the ADTF 100, quench probe 300, collection filter 210, and supplemental filter 220. Vacuum pump 230 is used to preferably provide a slight vacuum on the quench probe 200. The outlet flow of vacuum pump 230 is measured by vacuum outlet flowmeter 240, which provides gas sample to CO analyzer 271, CO ₂ analyzer 272, and O ₂ analyzer 273, the flows to each of which are measured by CO flowmeter 261, CO ₂ flowmeter 262, and O ₂ flowmeter 263, respectively. Excess gas is vented to atmosphere.

Fig. 5 presents a detailed design sketch of quench probe 300, which comprises tube 325 surrounded by alumina insulating cylinder 320 and metal shield 330. Quench port 310 provides for injection of cooling nitrogen into the hot exhaust from ADTF 100. Cooling nitrogen is fed to quench probe 300 via nitrogen inlet 350 into the water jacket 326 of tube 325. Cooling water enters cooling jacket 326 via water inlet 360 and exits via water outlet 365.

Referring now to Figs. 3 and 5, the residence time of particles within furnace tube 101 is controlled by inserting quench probe 300 into furnace tube 101 to a variable degree. The further that quench probe 300 is inserted into furnace tube 101, the shorter the transit distance of particles through furnace tube 101. Thus the low-density microspheres produced are quenched so as to trap the blowing agent in the microspheres, and solidified producing a dimensionally stable product. Quenching to 900 ⁰ C or below is preferably accomplished within 200 msec, more preferably within 120 msec, and most preferably within 60 msec. As will be apparent to those skilled in the art, quenching can be accomplished by means of nitrogen or other gas, water, a combination thereof, or other means known in the art without departing from the scope of the invention.

Referring again to Fig. 4, after quenching of the products exiting furnace tube 101 in quench probe 300, the quenched materials are collected at collection filter 210. Collection filter 210 may comprise a filter as is known in the art for collecting ash in bulk or any size-segregating devices known in the art such as a University of Washington Mark 5 source test cascade impactor

or a U.S. Environmental Protection Agency (EPA) Southern Research Institute five-stage multi- cyclone. Other collection devices as are well-known in the art ma}' of course be used without departing from the scope of the invention.

EXAMPLES To more clearly illustrate the present invention, several examples are presented below.

These examples are intended to be illustrative and no limitations to the present invention should be drawn or inferred from the examples presented herein. Low-Density Microspheres - Production Conditions

The production of strong microspheres from aluminosilicate materials is dependent upon the chemical and physical properties of the aluminosilicate precursor materials and the time- temperature history. It has now been discovered that the initial stages of cenosphere and low- density microsphere formation are influenced, at least in part, by the surface tension of the components in the formulation and their ability to wet each other and form a viscous liquid that will flow. Without intending to be bound by theory, it now appears that a driving force for the formation of low-density microspheres (including cenospheres) is thus the surface tension of the liquid phases. The viscosity at the temperature of formation drives the rate at which the liquids will flow. The following Examples will describe how this understanding may be used in the development of low-density microspheres and the conditions for their production in accordance with the present disclosure. Example 1 - Comparative

The following steps were followed in sequence to calculate the conditions for low-density microsphere formation from a formulation comprising flyash produced from combustion of a Texas lignite. The bulk flyash composition (wt-%) is provided below in Table 1.

Table 1

^" siόl Al ₂ O ₃ Fe ₂ O ₃ CaC) MgO SO ₃ ^" Na ₂ O K ₂ I) TiO ₂ ^" Mn ₂ O ₃ P ₂ O^ Tot 71.44 ϊTϋ 4~94 623 Tϊδ - (L30 L45 L02 - (U7 99.99

First, computer-controlled scanning electron microscopy (CCSEM) point count was performed on the feed ash. This provided composition information for a representative number of ash particles. Next, viscosity calculations were made at a specified temperature (1300 ⁰ C) on the CCSEM data using the Kalmanovitch-modified Urbain viscosity model. A determination of the number of feed ash particles having a calculated value for Iogi ₀ (viscosity in poise) of less than 3 (i.e., a viscosity under 1000 poise) was then made. The specified temperature was then adjusted,

and the viscosities of the ash particles were recalculated. This was done iteratively, until about 80 % of the flyash particles met the viscosity criterion.

Surface tension calculations were then made on the same data at the resulting temperature using the Al-Otoom model discussed above. The binder was then formulated such that the binder surface tension at the specified temperature was lower than the surface tension of the bulk of the feed flyash material at the same temperature. The resulting mixture was then prepared, a suitable blowing agent added to prepare a precursor, and the precursor was fired in the ADTF at the calculated temperature.

The insertion depth of the quench probe was then varied to vary the particle residence time and optimize the particle time-temperature history to optimize the yield of low-density microspheres. Fig. 6 illustrates the viscosity distribution histogram for the Texas lignite fly ash particles of this example calculated at 1300 ⁰ C. As can be seen from the histogram, this sample did not meet the viscosity criterion at 1300 ⁰ C and a poor yield was expected at this temperature. Example 2 - Calculating Operating Conditions The processes involved in the formation of synthetic cenospheres or microspheres are modeled based on the detailed understanding of ash and cenosphere formation in coal combustion systems. It has now been discovered that when the surface tension of the liquid phases is between about 295 and about 310 dynes/cm, the resulting wetting is conducive to the formation of spherical particles. The relationship between surface tension and temperature for an exemplary coal ash is shown in Figure 7, as calculated using the Al-Otoom model referred to above. The corresponding bulk ash composition (wt-%) is provided in Table 2 below.

Table 2

SiO ₂ Al ₂ O ₃ Fe ₂ O ₃ CaO MgO SO ₃ Na ₂ O K ₂ O TiO ₂ Mn ₂ O ₃ P ₂ O ₅ Tot

55.3 28.6 7.1 0.3 1.8 0.0 0.4 3.5 1.0 - 0.1 98.1

In addition, it has now been discovered that the viscosity of the molten liquid from which the low-density microspheres form should preferably be such that the value of log _]0 [μ (poise)] falls within the range of 2.5 to 3.0, more preferably within the range of 2.6 to 3.0, and most preferably within the range of 2.8 to 3.0 (inclusive), which is guided by the composition and temperature. A plot of the viscosity versus temperature for the above coal ash, calculated using the Kalmanovitch-modified Urbain model for the composition given in Table 2 is shown in Figure 8. Several measured viscosity points are also shown to illustrate the agreement between the model and actual viscosity.

This information has been used to develop formulations for precursor materials for low-

density microspheres, and is also useful in defining physical characteristics of the produced microspheres such as their density and strength.

As discussed above, the low-density microsphere precursor materials preferably comprise a binder (which may typically be a sodium silicate, though other similar materials may be used), an aluminosilicate material (including, without limitation, coarse fly ash, clay, and volcanic ash), and an optional blowing agent. The preferred characteristics for the binder are that it has a lower surface tension at temperature than the aluminosilicate material, and that it will assimilate the aluminosilicate in the melt. A comparison of the surface tension of a sodium silicate binder and the coal ash discussed above is shown in Figure 9. According to the physical model described by the above examples, during the formation of low-density microspheres, the lower surface tension binding material will rapidly move to, and become concentrated on, the surface of the forming microsphere, resulting in the sealing of the surface so that the blowing agent can react and expand the particle. As this is occurring, the aluminosilicate must be assimilating into the overall melt, resulting in the formation of a relatively homogeneous liquid phase. The viscosity of the resulting material will drive the rate of flow and the expansion of the particle, the thickness of the wall and, ultimately, the strength of the resulting microsphere.

The homogeneity and thickness of the microsphere shell depends upon the rate of sintering or combining of the particles forming the surface layer, and the resulting viscosity of the shell. The viscosity of this shell influences the rate of expansion during the reactions of the blowing agent.

As has been noted, the relationship between the viscosity and temperature largely determines the formation of low-density microspheres. The viscosity of the liquid should preferably remain within the range of about 2.6 to about 3.0 = [log ₁₀ (μ)], more preferably within the range of about 2.8 to about 3.0 = [logjo (μ)],where μ is expressed in units of poise, for a selected residence time in order to produce microspheres with desired wall thicknesses. The thickness of the microsphere wall influences the strength of the microsphere. A higher amount of water floaters often indicates a thinner microsphere wall. Example 3 - Synthetic Ash (Elite) A precursor formulation was prepared by combining 80 wt-% of illite ground to an average size of about 5 μm with 20 wt-% commercial grade sodium silicate solution (SiO ₂ ZNa ₂ O approximately 3.22, 40% solids content). A blowing agent comprising lignite-derived activated carbon was added to comprise 0.1 wt-% of the final precursor. The mixture was blended into a homogeneous slurry and dried at room temperature for about 5 minutes and at about 50 ⁰ C for about 20 hours, after which it was ground and sieved to obtain powder with a size range of from

about 180 m to about 425μm. Fig. 10 shows the viscosity-temperature profile for the illite sample, calculated according to the Kalmanovitch-modified Urbain (EERC) model described above. The calculated relationship between surface tension and temperature is also shown in Fig. 10 for both the illite sample and the sodium silicate binder. It should be noted that in accordance with the method of the invention, the surface tension of the binder is less than that of the illite. The surface tension of the illite is about 300-304 dyne/cm between 1350° and 1400 ⁰ C, and the viscosity of the illite is about 3 (logio(poise)) at about 1325°C, as shown in the expanded scale of Fig. 11. Accordingly, the sample was fired at a predetermined time-temperature history in the ADTF at an initial peak temperature of 1325°C and rapidly quenched as described previously. The ADTF feed rate was about 0.14 grams/minute. The ADTF gas flow was about 1 liter/minute primary air with about 3 liters/minute nitrogen secondary gas. The ADTF residence time was approximately 1 second. The low-density microspheres produced were rapidly quenched and collected. The distance from the injection point to the quench probe was 126 cm.

The bulk density of the resulting product was 1.21 g/cm ³ . Figure 12 is a scanning electron micrograph illustrating the low-density microspheres produced. Examples 4-9 - Silica Gel (1693-95, 1706-08)

A precursor formulation was prepared according to the present invention comprising approximately 75 wt-% silica gel and 25 wt-% sodium silicate with 0.1 wt-% sugar according to the following procedure. Laboratory grade silica gel was ground in a small hammermill to a nominal 1 mm particle size. 154 grams of sodium metasilicate (Na ₂ SiO ₃ .9H ₂ O) was dissolved in 200 ml distilled water. The solution was heated to approximately 38 ⁰ C to facilitate dissolving the sodium silicate. 0.269 grams of sugar (0.1% by weight of the anhydrous 25% sodium silicate + 75% silica gel) was added as a blowing agent. 200 grams of silica gel was slowly added to the sodium silicate solution with stirring. The silica gel was observed to adsorb all of the solution turning to a viscous gel that began to solidify. The mixture was oven dried at 110 ⁰ C for approximately 48 hours. The gel was found to have solidified into large very hard lumps, with no evidence of the previous grinding discernable. The mixture was broken into pieces able to be fed into a small hammer mill, where it was ground and the ground material sonic sieved to separate the desired particle size range. Oversize material was alternately reground and sieved. The size range of the precursor particles as tested was approximately 150 - 250 μm.

The calculated relationship between surface tension and temperature is shown in Fig. 13 for the silica gel. The surface tension of the sodium silicate binder calculated via the Al-Otoom model is slightly less than that of the silica gel over the temperature range shown (compare Fig 13

and Fig. 9). Fig. 14 shows the viscosity-temperature profile for the precursor particles, calculated according to the Kalmanovitch-modified Urbain (EERC) model described above. The surface tension of the silica gel is about 302-300 dyne/cm between 1350° and 1400 ⁰ C, and the viscosity of the precursor particles is such that logio(poise) is approximately 3.0 at about 1300 ⁰ C, as shown in Fig. 14.

Low-density microspheres were produced at two firing temperatures, 1325 ⁰ C and 135O ⁰ C, and the residence time of the precursor particles at the firing temperature was varied from about 0.7 to about 1.7 seconds. The results in terms of the bulk density of the resulting low- density microspheres are provided in Table 3 at the conditions described for Examples 4-9. Figure 15 is a plot of the bulk density of the produced microspheres as a function of drop height. Particle residence time is proportional to drop height, with 46.5 cm corresponding to a residence time of about 0.7 seconds.

Table 3

Example Temp ( ⁰ C) Drop ht (cm) Density (g/cm ³ ) Res time (sec)

4 1300 83.3 0.97 1.25

5 1300 112 1.14 1.7

6 1325 83.3 1.01 1.25

7 1300 46.5 0.73 0.7

8 1325 46.5 0.65 0.7

9 1325 62.5 0.62 0.94

Fig. 16 is a low-magnification scanning electron micrograph of low-density microspheres from Example 4 that are representative of those produced in these examples. Fig. 17 and Fig. 18 are higher magnification micrographs of individual microspheres produce in Examples 4 and 5, respectively. Example 10 - Silica Gel - High- Yield Water-Float (1710)

A precursor formulation was prepared as described in Example 4. The precursor was then fired as described in Examples 4-9 in a down-fired, entrained flow, drop tube furnace, and rapidly quenched. The peak temperature used was 1325°C, with a drop height of 62.5 cm, and a residence time of 0.94 sec, as shown in Table 4. Sufficient precursor was fired to produce quantities of low-density microspheres to enable reliable measurement of the water-float yield of the produced microspheres.

The bulk density of the low-density microspheres produced was 0.73 g/cm ³ . The fraction of water-float material produced was 0.8533. The water float yield of the microspheres produced was 85.33 wt-%. Table 4

Example Temp ( ⁰ C) Drop ht (cm) Density (g/cm ³ ) Res time (sec) Yield (wt-%)

10 1325 62.5 0.73 0.94 85.33

While the preferred embodiments of the invention have been shown and described, modifications thereof can be made by one skilled in the art without departing from the spirit and teachings of the invention. The embodiments described herein are exemplary only, and are not intended to be limiting. Many variations and modifications of the invention disclosed herein are possible and are within the scope of the invention. Accordingly, the scope of protection is not limited by the description set out above, but is only limited by the claims which follow, that scope including all equivalents of the subject matter of the claims.

The examples provided in the disclosure are presented for illustration and explanation purposes only and are not intended to limit the claims or embodiment of this invention. While the preferred embodiments of the invention have been shown and described, modifications thereof can be made by one skilled in the art without departing from the spirit and teachings of the invention. Process criteria, equipment, and the like for any given implementation of the invention will be readily ascertainable to one of skill in the art based upon the disclosure herein. The embodiments described herein are exemplary only, and are not intended to be limiting. Many variations and modifications of the invention disclosed herein are possible and are within the scope of the invention. Use of the term "optionally" with respect to any element of the invention is intended to

mean that the subject element is required, or alternatively, is not required. Both alternatives are intended to be within the scope of the invention.

The discussion of a reference in the Description of the Related Art is not an admission that it is prior art to the present invention, especially any reference that may have a publication date after the priority date of this application. The disclosures of all patents, patent applications, and publications cited herein are hereby incorporated herein by reference in their entirety, to the extent that they provide exemplary, procedural, or other details supplementary to those set forth herein.