Thomas W. Peterson

Metal capture by sorbents in combustion processes

Abstract The use of sorbents to control trace metal emissions from combustion processes was investigated, and the underlying mechanisms governing the interactions between trace metals and sorbents, were explored. Emphasis was on mechanisms in which the metal vapor was reactively scavenged by simple commercial sorbents, to form water unleachable products, which are easy to collect and isolate from the environment. Results are presented from two different scales of experimentation, involving a bench scale thermo-gravimetric reactor and a 17 kW down-fired laboratory combustor, respectively. Results from the bench scale tests showed that lead and cadmium, vaporized from the chloride salt, could be reactively captured at temperatures above the dew point. Both kaolinite and bauxite were effective sorbents for lead, while bauxite but not kaolinite was effective for cadmium. The primary reaction products, as identified by X-ray diffraction analyses, consisted of lead and cadmium aluminosilicates. Laboratory combustor tests, completed in the absence of coal char or coal ash particles, showed that lead could be effectively reactively scavenged in situ, in a combustor, downstream of the primary flame. Here, the high temperatures of the combustion process were being exploited to promote the reactions between the metal vapor and kaolinite sorbent, that were identified in the bench scale tests.

Similarity solutions for the population balance equation describing particle fragmentation

Similarity solutions are obtained to the population balance equation describing particulate systems undergoing fragmentation. The solutions are restricted only by the requirement that the breakage rate of particles of volume v be of the form φ(v) = Avb and the breakage distribution function of the form ω(u, v) = ƒ(v/u)/u, where u is the volume of the fragmenting particle. Under such conditions, the moments of the distribution asymptotically approach the form Ni(t) α t (1−i)/b , and the particle size distribution function is shown to obey a first-order linear ordinary integro-diflerential equation. For the case ƒ(v/u) = γ(v/u)γ−2, analytical solutions to the above equation were obtained. Complete solutions as well as asymptotic behavior are given. The results are potentially applicable to a wide range of particle fragmentation problems, including char/ash fragmentation during pulverized coal combustion, explosively generated aerosol formation, ore comminution, powder crushing and grinding, floe breakage, a...

The partitioning of iron during the combustion of pulverized coal

Abstract The partitioning of iron during pulverized coal combustion was investigated both experimentally and theoretically. Emphasis was on determining how coal variables and combustion conditions influenced the formation of slagging precursors. Experimental work consisted of burning a suite of six well-characterized coals in an aerodynamically well-defined 17 kW downflow combustor. Speciation of iron in collected ash samples was determined by Mossbauer Spectroscopy. A model was developed to predict the partitioning of iron for the coals and experimental conditions examined. The model was based on a competition between pyrite oxidation and iron capture by silicates and required as input, CCSEM data on the initial distribution of mineral matter in the coal. It used literature-derived mechanisms and kinetics to calculate pyrite oxidation and char combustion rates, together with a new glass formation mechanism, coupled with current knowledge of sintering. Requiring only two unknown parameters to be fitted, the model agreed well with 26 sets of experimental data, which covered a wide range of iron partitioning in the fly ash. Both experimental and theoretical results suggest that removal from the parent coal, of extraneous iron alone, may not be very effective in eliminating slag formation in boilers.

Modeling of multiphase atmospheric aerosols

This paper presents a mathematical model that describes the chemical composition and growth rate of atmospheric aerosols containing sulfate and nitrate, and that is applicable over the entire range of relative humidities. The model describes aerosols as consisting of an insoluble core, a soluble solid shell and an aqueous film. The chemical composition of the solid and aqueous aerosol phases in this model is governed by the following processes: 1. (1) the chemical equilibria between the ambient gas and the aerosol liquid phases and between the aerosol liquid and the soluble solid phases; 2. (2) the diffusion-limited condensation of H2SO4 and 3. (3) the liquid-phase chemical reaction of SO2. At low humidities the liquid phase may not be thermodynamically viable, in which case the model then treats the solid-phase/gas-phase equilibrium. Model simulations were conducted to study the effect of ambient relative humidity and gas-phase concentrations on the aerosol chemical composition and growth rate. Results indicate that 1. (1) precipitation of (NH4)2SO4 occurs at its point of deliquescence (i.e. at a relative humidity of close to 80%) even in the presence of other electrolytes; 2. (2) within the uncertainty of the thermodynamic data, the saturation products of NH3 and HNO3 for a liquid-coated aerosol and a dry aerosol are in agreement with each other; 3. (3) the aerosol concentrations correspond closely to (NH4) 2SO4 and NH4NO3 for both liquid and solid phases, i.e. NH4HSO4 is predicted to exist in negligible amounts under most conditions and 4. (4) the assumption of either an ideal internal mixture or an external mixture for dry sulfates and nitrate aerosols has little effect on model predictions.

Sodium partitioning in a pulverzed coal combustion environment

Sodium transformation mechanisms during pulverized coal combustion were investigated as functions of combustion temperature and of the modes in which sodium occurred in the parent coal. The combustion experiments took place in a 17-kW laboratory downflow combustor which was designed to form a link between bench-scale reactor studies and commercial-scale combustors. Three different coals were burned under excess air conditions, and size-segregated fly-ash samples were extracted far from the combustion zone. Atomic absorption on each ash sample yielded bulk composition data. Computer-controlled scanning electron spectroscopy (CCSEM) yielded composition data of individual ash particles and allowed inference of mechanisms governing the formation of these individual particles. It was demonstrated that partitioning of sodium in the vapor, and the subsequent fraction of sodium in the submicron fume, is greatly reduced by the presence of silicates and high temperatures. This was attributed to reaction, of sodium in the vapor with solid silicates. The CCSEM analysed showed that reaction to form sodium aluminum silicates was preferred over reactions forming sodium silicates, alone, although both are formed. Some physical condensation also occurs. Both included and excluded alumino-silicates are effective in reactively scavenging sodium. The maximum amount of sodium scavenged in any one particle was consistent with formation of Na 2 O·Al 2 O 3 ·2SiO 2 . It was demonstrated that sodium from coal reacted with kaolinite additives mixed with the coal, suggesting that this is a strategy to reactively scavenge vaporized sodium. Additional bench-scale experiments and equilibrium calculations suggest that the mechanism of sodium reaction with alumino-silicates involves sodium as sodium hydroxide.

Thermodynamics of multicomponent electrolytic aerosols

Abstract Considerable evidence exists for the presence of multicomponent aqueous electrolytic aerosols in the atmosphere. The size and chemical composition of these aerosols depend in large part on the interfacial equilibrium between the aerosol and surrounding gases, and correct description of condensation/evaporation on/from the particle requires knowledge of the thermodynamics at the interface. This work examines four models used for predicting the water activity and solute activity coefficients of electrolytic aerosols. The predictions of these models are compared to data of various electrolyte solutions, including data on H 2 SO 4 -(NH 4 ) 2 SO 4 -H 2 O and (NH 4 ) 2 SO 4 -NH 4 NO 3 -H 2 O systems. Particular emphasis is placed on model predictions for low water activities, which are important for aerosol modeling in arid regions.

Effect of Coal Type and Residence Time on the Submicron Aerosol Distribution from Pulverized Coal Combustion

Three types of pulverized coal were burned in a laboratory furnace under various combustion configurations. Pulverized samples of Utah bituminous, Beulah (North Dakota) lignite, and Texas lignite coals were burned at a rate of 2.5 kg/hr in a laboratory furnace. Aerosol size distributions were measured at various positions within the convection section, and temperature and gas compositions were measured throughout. The evolution of the submicron particle size distribution within the convection section for the three coals was similar, although the location of the initial particle mode at the convection section inlet varied with coal type. While staged combustion of Utah bituminous coal had a variable effect on the volume of submicron aerosol produced, staged combustion of the lignites caused a definite increase in the submicron aerosol volume. Vapor enhancement due to a localized reducing atmosphere, which would effect coals of higher ash volatility, is thought to explain this behavior.

ALKALI METAL PARTITIONING IN ASH FROM PULVERIZED COAL COMBUSTION

Abstract The relationship between the original form of sodium and potassium in coal and the ultimate fate of these metals was explored through experimentation in a 17kw laboratory downflow combustor. Composition analyses of size segregated samples of airbone ash at exhaust conditions for six different coals are reported. In all cases sodium was enriched in the small particle size range, and was shown both to form a sodium rich fume as well as an enriched surface layer around existing particles. The fraction of sodium appearing in the small particle size range, and therefore previously vaporized, could be correlated with the fraction of sodium that was acid soluble and inversely with the amount of silicon present in the parent coal. Increased temperature led to a decrease in the amount of sodium vaporized. For the coals examined, potassium was present primarily in the mineral form. Although the fraction of potassium vaporized could be positively correlated neither with total nor with acid soluble forms of ...

Hazardous waste incineration: The in-situ capture of lead by sorbents in a laboratory down-flow combustor

Experiments on a 17 KW downflow combustor determined how sorbent injection into the post flame influenced the particle size distribution of a lead aerosol formed from a surrogate lead containing waste. In the absence of chlorine, the lead aerosol size distribution evolved within the combustor to lie predominantly between 0.02 and 0.2 microns by the time it was sampled, at the combustor exit. When a commercial kaolinite sorbent was added, the lead sampled in that particle size range was reduced by 99%, and there is clear evidence that the heavy metal was reactively scavenged in the combustor by the larger sorbent particles. Chlorine kept the lead in the vapor form, until it was sampled, at which point it formed a fume in the probe. At large Cl/Pb ratios (greater than 10), the addition of sorbent was not effective in scavenging the lead vapor, although capture was again apparent as Cl/Pb ratios were reduced to 2. Data from experimental runs with a large excess of chlorine present, but not those where Cl/Pb ratio was equal to 2, are in sharp contrast to literature data from bench scale reactor studies.

Interactions between vapor-phase mercury compounds and coal char in synthetic flue gas

Data suggest that coal-fired power plants are a significant source of atmospheric mercury. Predicting emissions of mercury and the speciation of mercury in combustion emissions cannot be done without a fundamental understanding of the chemical reactions of mercury in flue gas. In this work, chars generated from three coals were used as sorbent material for both elemental mercury and mercuric chloride. The temperature of the source as well as the char sorbent was carefully controlled at either 343 or 433 K (70 or 160°C). When exposed to a synthetic flue gas consisting of O2, H2O, CO2, and N2, both Hg0 and HgCl2 were adsorbed by coal char. The rank of the coal seemed to have a large effect on the adsorption of Hg0, but not on adsorption of HgCl2. The bituminous chars adsorbed similar amounts of Hg0, while the sub-bituminous char adsorbed almost an order of magnitude less. The amount of Hg0 adsorbed did not appear to be correlated with the sulfur content of the char. The rank dependence suggests that some other characteristic of the char, for example, pore structure or surface functional groups may be important for adsorption of elemental mercury. Adsorption of HgCl2 was higher by a factor of two for the bituminous chars and 50 times higher for the sub-bituminous char. The adsorption of HgCl2 showed less differences as a function of coal rank and a better agreement among all chars with respect to char surface area. This suggests that adsorption of HgCl2 on char may be by a physical adsorption process and, therefore, char surface area would be a good indicator of capacity for HgCl2.