Search

Many high purity materials can be synthesized by a self-sustaining gas-phase flame for example: photovoltaic silicon; advanced fuels (e.g., boron); refractory metals (e.g., titanium, tantalum, zirconium, hafnium and niobium); and ceramics (e.g., silicon carbide, silicon nitride, tantalum carbide, titanium boride, and zirconium boride). Reactive metals, e.g., sodium or magnesium, combined with metal halides or mixtures of halides react hypergolically to produce the desired product as an aerosol and the by-product, reactive metal halide, in the gas phase. The key to the success of the process is the separation of the product from the by-product. This has been achieved by a novel supersonic impaction technique. The reaction products are expanded through a nozzle to produce a supersonic stream which is caused to impact a surface to produce a shock wave. The particles, unable to make the sharp turn behind the shock wave along with the gas, impact the surface where they are collected either as a solid, liquid, or powder, depending upon the surface configuration and temperature. The process has been examined thermodynamically for a large number of systems, and it has been demonstrated experimentally for silicon at 1.5 kg/h and yields exceeding 90%. Product purity is very high because of self-purification mechanisms inherent in the process.

The possibility of employing a gaseous self-sustaining reaction to prepare pure ceramic powders, particularly silicon nitride, for use in advanced ceramic materials has been explored. The thermodynamics of the process were first examined to determine possible operating conditions. Ignition experiments were then carried out in a closed bomb in which the reactant gases were silane and either hydrazine or a mixture of hydrazine and ammonia. Intense fuse wire ignition sources were used to determine the ignition limits over a broad range of initial conditions. These experiments were followed by flame propagation experiments through a tube into which the reactants were constantly fed and the flame repeatedly ignited at one end in order to produce product for preliminary determination of the powder properties. These experiments showed the feasibility of the flame synthesis process. A flat flame burner apparatus was then designed and constructed. A large number of experiments were carried out to demonstrate the regions of flame stability and to prepare product for evaluation. Flames were stable over a broad range of conditions: equivalence ratios from 0.1 to 1.1, unburned gas flow velocities from 20 to 150 cm/s and adiabatic flame temperatures from about 900 to 2000 K. High quality Si 3 N 4 powders have been prepared in this (small) laboratory flame apparatus at the rate of 10 g/min, and at yields of close to 100%.

Microstructural analysis was performed on several crystalline SiC samples previously prepared by three separate processing methods and subsequently sintered under high pressure and high temperature conditions using the cubic anvil cell MAX80 at Hasylab. Microcrystalline SiC was prepared using SHS conditions, while nanocrystalline SiC was prepared using both combustion synthesis methods and polymer precursors. High purity, highly disordered nanocrystalline SiC powders, with average particle diameters below 100 nm, were synthesized via combustion methods from precise mixtures of silane and acetylene. The properties of the silicon carbide powders prepared in this manner were dependent on the initial stoichiometry and pressure of the combustion mixture. Pyrolysis of polymer precursors to SiC was also used to fabricate ceramic powders containing uniformly-sized, highly disordered nanocrystalline SiC grains. The grain sizes ranged from approximately 3 nm to greater than 50 nm, and depended on the initial composition of the polymer, the pyrolysis conditions, as well as the annealing atmosphere, temperature and time. This paper describes the preparation methods for each of the SiC powders, the densification procedure, and preliminary results obtained primarily from transmission electron microscopy and X-ray diffraction analysis.

Silane/acetylene/ammonia flames were used to produce fine powders of composition SiCxNy The experiments were conducted in a spherical reactor at 1 atmosphere total pressure with spark ignition in the center of the chamber. The composition was determined from complete quantitative analysis of the gas products and elemental analysis of the solids. IR spectra indicate little bound hydrogen in the powders, but oxygen was adsorbed on the high surface area powders (60–80 m2/g, N2 BET). The experimental stoichiometries of the nascent powders can be represented as mixtures with weight ratios of Si3N4/SiC ranging from about 0.2 to 2, and variable amounts of excess carbon or silicon. The powders, from TEM, consist of chains of nearly symmetrical particles with diameters around 40 nm. Helium pycnometry densities in the range of 2.9 to 3.2 g/cc are consistent with the theoretical densities for the Si3N4/SiC stoichiometries. The powders were amorphous or nanocrystalline, from X-ray diffraction (Cu Kα). Annealing in argon at 1773 K for 8 h resulted in preferential growth of crystalline SiC domains.

Self-propagating flames in pure silane-acetylene mixtures produce silicon carbide (SiC) powder and hydrogen as main products. Through precise control of the stoichiometry of the reactant gas mixture, it has been possible to produce white SiC at high yields. Characterization of such powders by TEM has shown that the nascent powder consists of polycrystalline hexagonal plates with a narrow size distribution (40 ± 7 nm diameter). Infrared spectroscopy of powders indicate microcrystalline SiC and little bound hydrogen. Chemical analysis by the ANSI method showed the powder to be >96 wt % SiC with an impurity of silica (3.9 weight %) due to air exposure of the powder. Traces (0.1 to 0.2 weight %) of both free carbon and free silicon were found. Metal impurities detected by SIMS were typically low: less than 10 ppba for aluminum, sodium, titanium and vanadium. Boron was observed at 10 ppma. Like the oxygen, the boron impurities are probably associated with exposure of the powders to the atmosphere.

A nonequilibrium plasma jet has been used to deposit diamond films on a number of substrates, including silicon, silicon nitride, alumina, and molybdenum. Hydrogen is passed through a glow discharge and expanded through a supersonic nozzle to produce a highly nonequilibrium jet. Methane is added downstream of the nozzle, where it mixes and reacts with the nonequilibrium concentration of hydrogen atoms. The resulting supersonic jet strikes the substrate surface producing a high quality (determined by laser Raman spectrometry) adherent diamond film. Because of the low jet temperature, substrate cooling is unnecessary. Diamond deposition rates have exceeded 2 mg/kWh and I μm/h averaged over 16 cm2 area; good quality films prepared at substrate temperatures below 600 K. have been

Total ion concentration profiles were measured in laminar, premixed, low pressure, 2.7 kPa, acetylene-oxygen flames at equivalence ratios, , from 1.5 to 4.0. Peak total ion concentrations decrease with increasingup to and beyond the soot threshold (c=2.4). Near the soot threshold, a second peak appears in the ion profiles. This second peak contains some of the same ions, e.g., C13H9+, C15H11+, C18H11+, and C19H11+, as the first peak which is ascribed to chemi-ionization. The second peak reaches a minimum, at =2.9, well on the fuel rich side of the soot threshold, and then rises again with increase in equivalence ratio. These results refute the arguments against an ionic mechanism based on reported observations that the total ion concentration increases at the soot threshold. In a sooting, =3.0, flame the ion concentration profiles measured are in excellent agreement with measurements by others using molecular beam sampling techniques. The decay in the total ion concentration profile occurs simultaneously with an increase in soot concentration, consistent with the viewpoint that neutral soot particles are produced by ion recombination. The results support the chemi-ionization, ion-molecule mechanism of soot formation in flames.

Data in the literature have been organized to give a detailed description of the chemical sequences leading to soot formation in a premixed laminar flame. Neutral free radical, polyacetylene, and polycyclic aromatic hydrocarbon mechanisms are not consistent with this description. However, when chemiions are assumed to be the precursor on which the free radicals, polyacetylenes, and polycyclic aromatic hydrocarbons repeatedly add in fast ion-molecule reactions, a mechanism can be developed which is consistent with experiment. This mechanism should be tested further by careful observations of both ion and neutral species (including free radical) profiles in flames and by developing a computer model of the process to compare with experiment.

Mass spectrometric measurements of flame ion concentration profiles have been made in low pressure rich and sooting acetylene/oxygen and benzene/oxygen flames. As both flames are made richer and approach sooting the predominant ion C3H3+ is replaced by large positive aromatic ions with mass greater than 300. The profile of these ions corresponds with a previously identified class of “soot precursor” species. Similarities and differences between the ion spectra in the two flames are discussed along with possible reasons for dual maxima in individual ion concentration vs distance profiles observed in sooting flames. The data are consistent with an ionic mechanism of soot nucleation.

Mathematical model for thermal and chemi- ionization processes in turbulent nonequilibrium afterburning rocket exhausts plume, studying neutral and charged species distribution

Mass spectrometry and Langmuir probes have been used in a study of negative- and secondary positive-ion formation in low-pressure flames. Individual positive- and negative-ion profiles have been examined as a function of pressure and flame composition. Two mechanisms for secondary positive ion formation are proposed: The first consists of a series of charge-exchange reactions involving the primary ion (e.g., CHO + ) and stable neutral combustion intermediates: CHO + +CH 2 O→CH 2 OH + +CO CHO + +CH 3 OH→CH 3 OH 2 + +CO CHO + +CH 3 CHO→CH 3 CHOH + +CO CHO + +C 2 H 2 O→C 2 H 3 O + +CO. The plausibility of these reactions has been demonstrated by addition of CH 2 O, CH 3 OH, and CH 3 CHO to atomic diffusion flames. The second mechanism is one in which a primary ion undergoes a chain of oxidative degradation reactions which includes the observed ions as intermediates. Addition of N 2 and CO 2 to flames, together with the observed effects of varying flame composition, has led to the conclusion that, in addition to CHO + , an important primary ion is C 3 H 3 + . The proposed process responsible for the formation of C 3 H 3 + is: CH * +C 2 H 2 →C 3 H 3 + + e . Langmuir probe studies of a C 3 H 8 -air flame at 40 Torr have confirmed the importance of diffusional loss mechanisms in low-pressure flames and have led to the conclusion that the recombination coefficient α is temperature-independent. The measured values for the recombination coefficient and ambipolar diffusion coefficient are: α =2.4×10 −7 cm 3 /sec; D a =60 cm 2 /sec (referred to 1 Torr, 298°K). Taken together with values in the literature we conclude that α =2.4±0.4×10 −7 cm 3 /sec independent of temperature and pressure. Mass-spectrometric negative-ion profiles demonstrate, for the first time, the presence of significant quantities of negative ions in flames at 1 to 3 Torr. Dissociative attachment processes are the most probable formation route, and arguments are presented for C 2 − as the first negative ion formed via e +C 2 H 2 O * →C 2 − +H 2 O.

The status of our understanding of the-mechanisms of ion formation in flames is summarized, including not only a review of the literature but also some new proposals on cumulative and chemi-ionization. The high observed concentrations of ions in the combustion zone cannot be accounted for by thermal ionization of impurities, of equilibrium species existing at the flame temperature, or of non-equilibrium species. Ionization of carbon particles may account for the ions in high temperature-rich hydrocarbon flames, but experimental evidence is poor. This explanation is unsatisfactory for flames in lean mixtures. Ionization via translational energy or a simple extension of the high-energy tail of the Boltzmann curve is too slow. The most probable mechanism except for high temperature flames where thermal ionization predominates is cumulative excitation by processes such as CO(1II)+C2(1σ2+)→CO(1σ+)+C2++e− or chemi-ionization by processes such as CO(3II)+C2(1σ2+)→CO(1σ+)+C2++e− The most critical problem in this field is the actual identification of the ion.

It is shown how Langmuir probes can be used in flames to obtain not only the positive ion concentration but electron concentrations and electron temperatures. An internal method is presented for checking the results by comparing the wall potential calculated from the above three quantities with the observed wall potential. Satisfactory agreement between these two values in hydrocarbon air/or oxygen flames from 1.5 to 760 mm Hg increases the confidence in the use of Langmuir probes to obtain plasma profile properties in flames. The accuracy of positive ion and electron collision cross sections appears to be the major factor limiting the accuracy of the probe. Detailed results are presented for a number of flames. The positive ion concentration always exceeds the electron concentration indicating the formation of negative ions. Electron temperatures exceed the gas temperature and do not decay as rapidly as might be expected. A mass spectrometric technique for obtaining ion profiles of good spatial resolution is outlined and detailed profiles are presented for an acetylene–oxygen flame at 2.5 mm Hg. The ion CHO+; peaks ahead of C3H3+; which precedes H3O+;. Many other ions are observed to peak at about the same position in the flame as C3H3+;. There are still problems, however, with respect to interpreting the results in terms of the first ion produced from neutral species and the sequence of ion molecule reactions which follows.

Summary Experimental results for hydrocarbon-air or oxygen flames are presented at pressures down to 3 mm Hg. These are interpreted to yield ion recombination coefficients which are independent of pressure and equal to 2×10−7 cc/sec, i.e., the rate of ion recombination is second order with respect to the ion concentration. The rate of ion formation is proportional to the pressure squared so the formation process is second order. The importance of ambipolar diffusion at pressures below 66 mm Hg is deduced, so it is necessary to consider diffusion in interpreting ion decay processes behind the flame front. These results with other available information are discussed in terms of ion production by chemi-ionization, e.g. by CH+O→CHO++e−, At 66 mm Hg, if the reactants are present in equal concentration, the reactant concentration would have to be 1011 molecules/cc. The initially produced ion rapidly reacts with H2O by an ionmolecule charge exchange reaction to produce H2O+. The half-life of the initially produced ion is very short, approximately 10−7 sec, so it is present in only small concentrations. In summary, a picture has been formulated of the mechanism of ion formation and removal in flames which is consistent with the known facts. Some details are still fuzzy, and the specific elementary steps to be chosen are still open to question. In such a complex system as a flame, with many reactions occurring, it is surpring that such an apparently simple picture seems to fit the experimental observations. With more accurate and detailed data, additional facets of the problem can be expected to appear.

The objective of this program is to determine the feasibility of using a non-equilibrium hydrogen plasma jet as a chemical synthesis tool in helping to meet the objectives of the JPL Low-Cost Silicon Solar Array Project. Reactions of hydrogen atoms (produced by a glow discharge and expanded through a nozzle) with chlorosilanes are being studied. During this period, previous results were evaluated and four possible processes identified for further study: (1) production of polycrystalline silicon photovoltaic surfaces, (2) production of SiHCl/sub 3/ from SiCl/sub 4/, (3) production of SiH/sub 4/ from SiHCl/sub 3/, and (4) purification of SiCl/sub 4/ by metal impurity nucleation. Estimates, based on homogeneous and wall recombination rates, indicate that the hydrogen atom concentration far downstream of the nozzle at the liquid nitrogen traps is less than 20% of its original value. This could account for the low yields, about 10% of SiCl/sub 4/ converted to SiHCl/sub 3/ collected in the traps, but does not explain the failure to observe reactions of SiCl/sub 4/ in the gas phase. In situ mass spectrometric measurements using SiHCl/sub 3/ indicated about 30% SiHCl/sub 3/ consumption; SiH/sub 4/ was qualitatively identified as a gas-phase product. Previous difficulties with the mass spectrometermore » have been corrected and these questions can now be addressed more accurately. The most striking result was the recognition that the strongly adhering silicon films, amorphous or polycrystalline, produced in the studies could be the basis for preparing a photovoltaic surface directly; this process has potential advantages over other vapor deposition processes.« less

Smoke related performance of both jet engine and research combustors has been correlated with several fuel properties. The smoke related data included: smoke number, liner temperature rise, and radiation flux to the combustor wall; fuel parameters included: percent hydrogen, percent aromatic, percent polycyclic aromatic, smoke point, and the threshold soot index, TSI. The research combustor results correlated best with the threshold sooting index. While some correlations with engine performance were excellent, no single fuel property was generally useful in evaluating smoke related performance, mainly because of insufficient data on the fuels tested in the jet engine programs, e.g., percent aromatics specifies a class of fuels which span > 50% of the possible range of sooting tendencies. It is, however, demonstrated that fuel composition plays a dominant role in determining smoke related engine parameters. It is recommended that fuels used for engine testing programs be chemically analyzed in greater detail or be made available for laboratory measurements of soot thresholds and soot yields until a sufficient data base is available to establish a laboratory technique of predicting relative smoke related performance of fuels in jet engines.

The original objective of this program was to determine the feasibility of high volume, low-cost production of high purity silane or solar cell grade silicon using a nonequilibrium plasma jet. The emphasis was changed near the end of the program to determine the feasibility of preparing photovoltaic amorphous silicon films directly using this method. The nonequilibrium plasma jet is produced by partially dissociating hydrogen to hydrogen atoms in a 50 to 100 Torr glow discharge and expanding the H/H/sub 2/ mixture through a nozzle. A high flux density of hydrogen atoms is thus produced at concentrations of about 3 mol % with about 30% energy utilization efficiency. The jet is mixed with a second reactant and the reaction proceeds at a temperature of 400 to 600/sup 0/K to produce products. Yields of SiH/sub 4/, SiHCl/sub 3/, or SiH/sub 2/Cl/sub 2/ from SiCl/sub 4/ and SiHCl/sub 3/ were too low to be economically attractive. However, both amorphous and polycrystalline silicon films which strongly adhered to Pyrex, Vycor, aluminum, or carbon were prepared with either SiCl/sub 4/ or SiHCl/sub 3/ reactants. Preliminary doping experiments with PH/sub 3/ did not alter the electrical resistivity of these films. Strongly adhering films with SiH/sub 4/more » reactant were more difficult to prepare; they were prepared by carefully cleaning the aluminum substrate, diluting the SiH/sub 4/ with about 90% argon, and forming the glow discharge between the mixing nozzle and the aluminum substrate. Doping such films with P by adding PH/sub 3/ reduced the electrical resistivity by two orders of magnitude. The nonequilibrium plasma jet should be further evaluated as a technique for producing high efficiency photovoltaic amorphous silicon films.« less

The feasibility of high volume, low cost production of high purity silane or solar grade silicon using a non-equilibrium hydrogen atom plasma jet was determined. Reactions of hydrogen atoms in the plasma jet with chlorosilanes (added either to the discharge or to the hydrogen atom stream) are being studied. Products of the H/SiCl/sub 4/ reaction were collected and quantitatively analyzed using a fractionation/distillation-mass spectrometric technique. Conversion of SiCl/sub 4/ to lower chlorosilanes and solid (Si) was on the order of 10 percent. The percentage conversion was found to depend upon equivalence ratio (ratio of SiCl/sub 4/ to H concentrations) and reaction zone pressure. Electronic grade reagent SiCl/sub 4/ gave smaller product yield and different product distribution than did technical grade. Real-time, in situ mass spectrometric sampling of the H/H/sub 2//SiCl/sub 4/ reaction zone gases was initiated. Quantitative results have not been obtained due to the complexity of the chlorosilane mass spectra, and consequent resolution and sensitivity difficulties with the mass spectrometer. Improvements in the instrumentation needed for these measurements are discussed. The performance of the flower petal nozzle, No. 5 was characterized. Current/voltage behavior was similar to that previously observed in the nozzle-as-anode mode of operation for simpler nozzles. Hydrogenmore » atom yields of approximately 1 percent are obtainable with production efficiencies of approximately 2 to 3 moles H/kWh of discharge energy. A remeasurement of yields and production efficiencies from the simpler nozzle No. 2 confirmed yields of approximately 3 percent for H atoms and production efficiencies of approximately 5 moles H/kWh. Measurements of H-atom yields were found to depend upon reagent mixing conditions, indicating the need to continue investigations of the mixing properties.« less

The objective of this program is to determine the feasibility of high volume, low-cost production of high purity silane or solar grade silicon using a non-equilibrium hydrogen plasma jet. Reactions of hydrogen atoms produced in the plasma jet with chlorosilanes (added either to the discharge or to the hydrogen atom stream) are being studied. During this quarter, hydrogen atom production in the plasma jet (the sine qua non for carrying out the chemical experiments) has been investigated with 3.5 and 6.1 mm diam nozzles both as anodes and as cathodes for the glow discharge. Complete energy balance data for the discharges have been obtained at H/sub 2/ pressures of 25, 50, and 225 Torr with emphasis on the 50 Torr data. Results were sufficiently encouraging (e.g., 2.7% (vol) H atom concentrations were obtained at 30% energy conversion into H atoms) that chemical synthesis studies were begun. SiCl/sub 4/ was initially added to the discharge with H/sub 2/ and metallic deposits were formed in the discharge region. Downstream, a brown powder was formed on the reactor walls. In further work, SiCl/sub 4/ was added to the H/H/sub 2/ jet. A reaction ensued, depositing a brown powder on the reactor walls. Qualitativemore » analytical tests of this powder are consistent with its being amorphous Si. The H/H/sub 2//SiCl/sub 4/ reaction results in a green chemiluminescence.« less

The objective of this program is to determine the feasibility of high volume, low-cost production of high purity silane or solar grade silicon using a non-equilibrium hydrogen atom plasma jet. Reactions of hydrogen atoms in the plasma jet with chlorosilanes (added either to the discharge or to the hydrogen atom stream) are being studied. Calculations were made to define the experimental limits within which gas-phase reactions would be most likely to occur. These show that H-atom recombination does not compete significantly with reaction as a sink for H, and that, if the H/SiCl/sub 4/ overall reaction has an activation energy, the pressure ratio across the nozzle and the temperature in the discharge become very important parameters. Most importantly, the maximum operating pressure is shown to be limited by the pressure at which the glow discharge can be maintained or the pressure at which the hydrogen atom yields become small. Solid products of the H/H/sub 2//SiCl/sub 4/ reaction were recovered. Depending upon experimental conditions, these were (tentatively) identified as a polymeric material, amorphous silicon and polycrystalline silicon. The polycrystalline material was shown by X-ray fluorescence analysis to have ppm levels of nickel present, but conspicuously absent at the ppm level weremore » the heavy metals Fe, Cr, Pb, Sn, Cu, and Zn, materials of construction of the apparatus.« less

Originally, studies of ionization in flames were motivated by the observation that in hydrocarbon flames, the ion concentration far exceeds the value expected if ionization were due to thermal processes alone(1,2) (see Table I). The objectives of these studies were to explain the source of nonequilibrium ionization and to explore links between flame ionization and the mechanism of flame propagation. An explanation of the source of flame ions was found in the process of chemi-ionization.(2) This led to further studies of the details of ionic reactions which occur in flames and of flame reactions which can be induced by the addition of foreign elements.