Reduction of Nitric Oxide from Combustion Flue Gas by Bituminous

May 3, 2003 - A detailed parametric study quantifying the effect of various operating parameters on the reduction of nitric oxide (NO) in the presence...
10 downloads 9 Views 168KB Size
2536

Ind. Eng. Chem. Res. 2003, 42, 2536-2543

Reduction of Nitric Oxide from Combustion Flue Gas by Bituminous Coal Char in the Presence of Oxygen Himanshu Gupta and Liang-Shih Fan* Department of Chemical Engineering, 121 Koffolt Laboratories, The Ohio State University, 140 West 19th Avenue, Columbus, Ohio 43210

A detailed parametric study quantifying the effect of various operating parameters on the reduction of nitric oxide (NO) in the presence of oxygen by inexpensive carbons was carried out. The selectivity of the carbon-NO reaction over the parasitic carbon-oxygen reaction was measured by a global selectivity parameter, defined as milligrams of NO reduced per gram of carbon consumed. An increase in the reaction temperature, bed height, and inlet NO concentration leads to a higher selectivity factor. However, there exists an optimum particle size of carbon that maximizes this selectivity. Whereas, in the absence of oxygen, the rate of the carbon-NO reaction was insignificant, the presence of oxygen in the inlet creates active sites that convert NO to nitrogen. An optimum inlet oxygen concentration provided the highest degree of NO reduction. The operating conditions can be selected to achieve >90% NO reduction at a selectivity of 60-100 mg of NO reduced/g of carbon. Introduction Numerous processes have been developed and commercialized to reduce nitric oxide (NO) emissions existing in flue gas due to fossil fuel combustion.1 Hightemperature primary measures, implemented in a coalfired boiler, include combustion modifications such as air staging, fuel staging, flue gas recirculation, and the use of low-NOx burners.2 Medium-temperature postcombustion NO reduction processes include selective noncatalytic NO reduction3 (SNCR), which operate at about 850-1000 °C, and selective catalytic reduction4 (SCR), which reduce NO in the 300-450 °C range. Lowtemperature processes employ NO oxidation by ozone to form nitrogen dioxide (NO2), followed by scrubbing of NO2 in basic slurries.5 Carbon-based processes have been studied for the reduction of NO emissions from stationary fossil-fuel combustion sources. In the reburning process, the injection of micronized coal over the flame leads to the creation of a reducing atmosphere in which hydrocarbon radicals reduce NO to nitrogen.6-8 Sub-bituminous char adsorbs NO and NO2 (NOx) in the presence of oxygen and moisture at room temperature.9 Combined SOx/NOx processes employ carbon as a catalyst for the reduction of NO by ammonia below 200 °C.10,11 Numerous studies have also been carried out to elucidate the mechanism of the various heterogeneous interactions (low-temperature physisorption, higher temperature chemisorption, and gasification) between carbon and NO. In this process, carbon reduces NOx in a lower temperature range (300-700 °C) in the presence of oxygen (1-5%). Below ambient temperature, NO has been shown to adsorb physically and reversibly on graphite.12 The extent of physisorption decreases with an increase in the temperature above ambient conditions. It was previously observed by us that the amount of physisorbed NO on a thermally pretreated activated carbon * To whom correspondence may be addressed. Telephone: (614)-292-7907. Fax: (614)-292-3769. E-mail: [email protected].

surface decreases with an increase in the temperature in the 25-215 °C range.13 A combination of physisorption and chemisorption was observed on a thermally cleaned phenol-formaldehyde-derived char in the 50150 °C temperature range.14 It has been postulated that while physisorption leads to the formation of C(NO)type surface species that evolve NO upon thermal desorption, chemisorption occurs by the dissociation of NO molecules to yield C(N)-, C(O)-, and C(O2)-type complexes on the carbon surface. The complexes were characterized based on the evolution of N2, CO, and CO2 upon thermal desorption, respectively.15 They showed that below 200 °C NO dissociatively chemisorbs on the carbon surface, yielding nitrogen molecules. Oxygen originating from the dissociatively chemisorbing NO molecules forms complexes with the carbon atoms on the surface, while nitrogen atoms combine in pairs and escape as nitrogen molecules.16 These complexes eventually saturate the surface of carbon, thereby limiting any further chemisorption of NO due to nonavailability of free carbon sites. As the temperature of the bed increases beyond 650-700 °C depending on the nature of carbon, these carbon-oxygen complexes break down to yield gaseous CO and CO2 and lead to the creation of new active sites on the surface, which in turn sustain the chemisorption of NO on a continued basis. The desorption of these complexes is an activated process and increases with increasing temperature. When the rate of desorption of complexes exceeds the rate at which all of the NO molecules react with the carbon surface, the NO reduction is essentially complete. The two-step mechanism has been widely accepted as the mechanism of the high-temperature carbon-NO reaction.17 The activation energy of the high-temperature desorption of complexes (28-59 kcal/mol) was usually much higher than that involved in the low-temperature dissociative chemisorption of NO (4.5-19 kcal/mol).18 In the higher temperature range where sustained NO reduction can take place without the limitation of the buildup of carbon-oxygen complexes on the surface, the

10.1021/ie020693n CCC: $25.00 © 2003 American Chemical Society Published on Web 05/03/2003

Ind. Eng. Chem. Res., Vol. 42, No. 12, 2003 2537

reaction between NO and carbon would lead to the formation of N2, CO, and CO2 as shown below:

C + 2NO f N2 + CO2

(1)

2C + 2NO f N2 + 2CO

(2)

The effect of oxygen on the carbon-NO reaction is of paramount significance because, in the flue gas, the oxygen concentration usually varies between 1 and 5%. This is much higher than the usual NO concentration of 250-1500 ppm. Compounded on the adverse concentration ratio is the higher oxidizing power of oxygen that would lead to enhanced parasitic consumption of carbon.19 Besides oxygen, side reactions could also occur with moisture and CO2 that are present in 2-7 and 5-16 vol % in flue gas. The main detrimental reactions are

C + O2 f CO2

(3)

C + 1/2O2 f CO

(4)

C + H2O f H2 + CO

(5)

C + CO2 f 2CO

(6)

However, the presence of oxygen in the range of 0.12% reduces the temperature required for sustained NO reduction. This occurs by the low-temperature gasification of carbon by oxygen, leading to the creation of active sites.20 These active sites, in turn reduce NO to nitrogen. The majority of the carbonaceous reductant is consumed by oxygen. The primary challenge in commercializing this carbon-based NO reduction process involves reducing the consumption of carbon by improving the selectivity of the carbon-NO reaction in the face of the other competing reactions. It has been shown that the rate of the carbon-NO reaction, although lower than the rate of the carbon-oxygen reaction, is much higher than the reactivity of the carbon-CO2 and carbon-H2O reactions.21 Hence, this study focuses only on the parasitic effect of oxygen. The motivation behind this study was to elucidate the effect of operating conditions on the degree of NO reduction by inexpensive carbonaceous sorbents that can be used on a consumable basis. Because we have conducted isothermal experiments, a global parameter for selectivity (milligrams of NO reduced per gram of carbon) has been used to quantify the effect of various operating parameters. Experimental Section Reactor Assembly. The experiments were carried out in a 2-in. stainless steel reactor tube housed in a Thermolyne 21100 tube furnace, as shown in Figure 1. Helium, NO in helium, and oxygen in helium were metered by variable-area flowmeters. The gaseous mixture was sent through the fixed bed of carbon in the reactor, and then gas analysis was conducted. The sorbent bed support consists of a 3/8-in.-o.d. steel tube with a perforated stainless steel disk support. The inlet gases were preheated in the annular zone of the 2-in. tube, after which the gases entered the sorbent bed. The gases exiting from the sorbent bed are analyzed for NO and oxygen. A chemiluminescence NOx analyzer (Advanced Pollution Inc. model 200 AH) provided the online measurement of the NO concentration. Oxygen was

Figure 1. Schematic diagram of the reactor setup used for conducting the integral experiments.

monitored using a Teledyne model 3000 PA percent oxygen analyzer that generates an electric current as a function of the oxygen concentration through its electrochemical galvanic micro fuel cell. Synthesis of Char. Eastern bituminous coal was used as the source material for the synthesis of char. Coal char was obtained by heating a bed of coal at 950 °C for more than 2 h in flowing inert nitrogen. The coal particles went through a plastic state and yielded a lump of char. The char (coal devoid of volatile matter) was then allowed to cool, crushed, and sieved to required sizes. The recovery of char from the original coal was about 48 wt %. Experiments were also carried out on a commercially available activated carbon called SXO (obtained from Carbon Corp., Columbus, OH). Experimental Procedure and Data Analysis. In these experiments, a fixed bed of char was brought to the reaction temperature in helium. The flow of gases was switched according to predetermined concentrations of NO and oxygen in helium. Outlet concentrations of NO and oxygen were continuously logged. The experiments were carried out for an hour, after which the flow of reactive gases was stopped and the char was allowed to cool in helium. The remaining char was weighed, and the mass of carbon consumed was evaluated from the difference between the initial and final weights of char. Integration of the area under the outlet NO concentration versus time profile yielded the amount of NO unreacted. The amount of NO reduced was estimated by the difference between the NO in the inlet and the NO passing through the bed unreacted. It is obvious that these experiments are integral in nature and the sorbent mass and physical properties evolve with time. The data are not amenable for kinetic analysis, but the effect of each operating variable can be quantified by a global selectivity factor, defined as milligrams of NO reduced per g of carbon consumed. Each of the parameters under consideration has two plots: one for

2538

Ind. Eng. Chem. Res., Vol. 42, No. 12, 2003

Figure 2. TPR of NO on bituminous coal char in the presence of oxygen (flow, 200 mL/min; inlet NO concentration, 1209 ppm; inlet oxygen concentration, 0.85%; wt of carbon, 300 mg).

the outlet NO concentration as a function of time and the other one depicting the variation in selectivity. Results and Discussion The parameters under consideration in this study include the reaction temperature, the char loading, inlet NO and oxygen concentrations, and the particle size of char. The following sections detail the influence of each of these parameters on the extent of NO reduction and the selectivity for the carbon-NO reaction. Effect of the Temperature. Oxygen is a stronger oxidizing agent than NO. Moreover, while it is evident that sustained reduction of NO requires over 650 °C for bituminous coal char, gasification of char in the presence of oxygen could occur at lower temperatures. A temperature-programmed reduction (TPR) was conducted on bituminous coal char to identify the temperature window in which a high extent of NO reduction could also occur in the presence of oxygen. This was achieved by continuously monitoring the outlet NO concentration while ramping up the carbon bed temperature. In an earlier study, we noticed that the char-NO reaction, conducted in the absence of oxygen, began only beyond 350 °C.22 In contrast, TPR results carried out in the presence of oxygen indicate the onset of NO reduction above 275 °C, as shown in Figure 2. With an increase in the temperature, the reaction rate increases, leading to a fall in the outlet NO concentration. This interaction continues for a while until the carbon-oxygen complexes completely cover the surface of char. Hence, the char is unable to offer any sites for continued dissociative chemisorption of NO. This leads to the escape of NO though the carbon bed remained unreacted, as seen Figure 2. As the temperature increases above 550 °C, the mechanism of this heterogeneous reaction changes, as mentioned earlier. Higher temperatures induce the thermal desorption of the carbon-oxygen complexes in the form of CO and CO2, leading to the creation of active sites. The NO reduction rate is now controlled by the rate of desorption of these complexes. These active sites provide new sites for continued NO reduction. The thermal desorption process is an activated phenomenon, and higher temperatures lead to an enhanced rate of desorption of complexes. Ultimately, from Figure 2, it can be observed that, under the process conditions used, the rate of desorption of these complexes is enough to reduce a majority of the NO molecules beyond 675 °C.

Figure 3. Effect of the temperature on the reduction of NO on SXO-activated carbon (flow, 200 mL/min; inlet NO concentration, 1300 ppm; wt of carbon, 150 mg; inlet oxygen concentration, 1.45%).

Hence, we chose a temperature range of 620-820 °C to carry out extended isothermal experiments. Previous studies that show a negligible reaction rate between char/activated carbons for NO reduction at 400 °C in the presence of oxygen also corroborate this observation on the active reaction temperature range.20,23 The effect of temperature on the reduction of NO by a commercially available activated carbon (tradename SXO; Carbon Corp., Columbus, OH) is shown in Figure 3. This carbon is characterized by a high initial nitrogen Brunauer-Emmett-Teller (BET) surface area of 818 m2/g and a pore volume of 0.49 mL/g. Activated carbon was chosen because of its high surface area and high potential reactivity. The outlet NO concentration for the 700 °C experimental run initially is almost zero, indicating almost complete NO reduction. Initially, the surface area contained in the entire carbon bed is capable of generating a sufficient number of active sites (by the gasification of carbon by O2 and NO) to react with all of the NO molecules flowing past the carbon bed. This continues for about 9-10 min. After this, the mass of carbon reduces because of its continued consumption by the gasification reactions, and subsequently the number of free sites created on the surface of the remaining carbon bed is insufficient for complete removal of all of the NO molecules, leading to the passage of unreacted NO through the bed. The rapid loss of carbon (due to higher gasification rates at 700 °C) causes a steep rise in the outlet. After 90 min, the entire carbon bed was gasified and only inert ash remained. However, at a lower temperature of 650 °C, the gasification of carbon is relatively slower, leading to a lower rate of generation of active sites on the surface. The active sites so formed are insufficient to reduce all of the NO molecules passing through the bed, leading to a substantial NO concentration in the outlet stream. With a further reduction in the reaction temperature, the gasification rate is even lower. Hence, the carbon bed lasts for a longer time, as proved by the

Ind. Eng. Chem. Res., Vol. 42, No. 12, 2003 2539

Figure 5. Effect of the temperature on the selectivity of bituminous coal char for the char-NO reaction (flow, 1000 mL/min; inlet NO concentration, 1000 ppm; wt of char, 1000 mg; inlet oxygen concentration, 2.0%). Figure 4. Effect of the temperature on the reduction of NO on bituminous coal char (flow, 1000 mL/min; inlet NO concentration, 1000 ppm; wt of char, 1000 mg; inlet oxygen concentration, 2.0%).

outlet NO concentrationprofile at 550 °C in Figure 3. It can be concluded that higher temperatures lead to a higher extent of NO reduction due to higher reaction rate. Data reduction shows that high-temperature operation required a lower consumption of carbon, with the selectivity being 79.11, 84.89, and 96.40 mg of NO/g of carbon for 550, 650, and 700 °C, respectively. Similarly, the effect of temperature was also quantified on the performance of bituminous coal char, the main sorbent under evaluation in this study. Unlike the SXO-activated carbon described above, the coal char has very low surface area and pore volume (0.96 m2/g and 0.000 56 mL/g, respectively). An oxygen concentration of 2.0% was chosen for this set of experiments in accordance with the oxygen concentration present in the flue gas at higher temperatures. Figure 4 shows that initially the total surface area in the entire char bed is not high enough to generate the requisite number of active sites to reduce completely all of the NO molecules entering the reactor. Hence, NO escapes from the char bed, as seen at all four temperatures (620-820 °C). However, as the reaction progresses, char becomes increasingly porous, creating a higher specific and total surface area in the bed. This increase in the surface area enhances char-oxygen and char-NO reactions, leading to the formation of more active sites. This, in turn, leads to higher NO reduction, and the outlet NO concentration drops, as seen at all four temperatures. Sustained reduction then leads to continued char consumption, and ultimately NO escapes through the char bed unreacted because there is not enough char for NO reduction. Figure 5 depicts the selectivity as a function of the reaction temperature. The selectivity increases from about 56 mg of NO/g of carbon at 620 °C to 102 mg of NO/g of carbon at 820 °C, as shown in Figure 5. The rate of desorption of the carbon-oxygen complexes increases with temperature, creating a higher concentration of active sites, on the surface of the remaining carbon bed, that can reduce NO molecules. Higher temperatures lead to the reaction between CO2 and carbon-forming CO. This CO so formed can also react with NO, aided by the catalytic effect of char and its constituents, thereby yielding a higher selectivity toward the carbon-NO reaction with rising reaction temperature.

Figure 6. Effect of the inlet oxygen concentration on the reduction of NO on bituminous coal char (flow, 1000 mL/min; inlet NO concentration, 1000 ppm; wt of char, 1000 mg; reaction temperature, 670 °C).

Effect of the Oxygen Concentration. The next parameter to be studied was the oxygen concentration. A range of 0-3% was chosen for this set of experiments to mimic the concentration of oxygen existing in the combustion flue gas at high temperature. A lower temperature of 670 °C was used to capture the outlet NO trends. The use of a higher temperature would lead to complete reduction of NO, and at lower temperatures, NO would not reduce to a sufficient degree, causing inaccurate data analysis due to the amplification of errors that are concomitant with low degree of NO reduction. The results of the effect of varying the inlet oxygen concentration are presented in Figure 6. In the absence of oxygen, it was observed earlier that virtually all of the NO passed through the char bed unreacted. However, it is evident from the figure that even a low concentration of oxygen (0.2%) leads to a measurable amount of NO reduction. As the oxygen concentration increases, so does the degree of NO reduction. This trend continues until about 1% oxygen concentration. Under the particular set of operating conditions used in this experimental set, it was observed that the highest degree of NO reduction occurred at 1% oxygen concentration. As the inlet oxygen concentration increases

2540

Ind. Eng. Chem. Res., Vol. 42, No. 12, 2003

Figure 7. Effect of the inlet oxygen concentration on the selectivity of the reaction between bituminous coal char for the char-NO reaction (flow, 1000 mL/min; inlet NO concentration, 1000 ppm; wt of char, 1000 mg; reaction temperature, 670 °C).

further, the overall rate of the carbon-oxygen reaction increases and a lower fraction of the active sites are available for NO reduction. The data in Figure 6 show that the char gets consumed very rapidly at 3% inlet oxygen concentration, leading to comparatively lower NO reduction over the course of the experiment. Selectivity of char as a function of the inlet oxygen concentration is shown in Figure 7. The figure shows an expected monotonic decrease with an increase in the inlet oxygen concentration. The selectivity drops from 220 to about 30 mg of NO/g of char reduced as the concentration of oxygen increases from 0.2% to 3%. In the absence of oxygen, carbon can react with only NO and therefore leads to the highest selectivity possible. The rate of the carbon-NO reaction in the absence of oxygen is low, and significantly higher temperatures would be needed to reduce a similar amount of NO under the process conditions used in this set. It is also obvious that the scavenging consumption of carbon by oxygen cannot be contained. However, it is clear that, other factors remaining the same, an “optimum” concentration of oxygen maximizes the extent of NO reduction. The presence of oxygen is therefore necessary to create the requisite number of active sites to reduce the maximum amount of NO. However, as the oxygen concentration exceeds this optimum value, there is a competition between NO and the remaining oxygen for these sites and the carbon-oxygen reaction prevails because of the higher reactivity of oxygen, leading to a loss in selectivity. Effect of the NO Concentration. The inlet NO concentration was varied in the 250-1250 ppm range to represent NO emissions from a host of coal combustion boiler types. The results obtained at 2% inlet oxygen and 670 °C are depicted in Figure 8. It can be concluded that outlet NO concentration decreases with lower inlet NO concentrations. This is in accordance with general reaction kinetics’ principles. However, the selectivity of char increases with an increase in the inlet NO concentration, as shown in Figure 9. Once again, this is due to the fact that once all of the oxygen is consumed in the reaction bed, the extent of NO reduction is now solely dependent on the availability of NO to the carbon surface. If there is an insufficient amount of NO (when the inlet NO concentration is low), then only a small amount of NO gets reduced. Conversely, a higher inlet NO concentration funnels more NO mol-

Figure 8. Effect of the inlet NO concentration on the isothermal reduction of NO on bituminous coal char (flow, 1000 mL/min; wt of carbon, 1000 mg; inlet oxygen concentration, 2.0%; temperature, 670 °C).

Figure 9. Effect of the inlet NO concentration on the selectivity of bituminous coal char for the char-NO reaction (flow, 1000 mL/ min; wt of carbon, 1000 mg; inlet oxygen concentration, 2.0%; temperature, 670 °C).

ecules to the carbon surface for reduction. Irrespective of the concentration of NO, the char consumption remains almost the same because it is governed by the consumption of oxygen. Thus, the selectivity increases with the inlet NO concentration. This set of data also signifies that the process is not stoichiometric with respect to the mass of carbon required for a particular reduction in the amount of NO. This is quite different from SNCR and SCR processes, where the amount of NO reduced follows the stoichiometric ratio between ammonia and NO accurately. From the data it is obvious that an identical amount of char under identical process conditions can reduce differing amounts of NO based on varying the inlet NO concentration. This could make this system more applicable to combustion systems that generate high NO emissions such as cyclone and cell burners. Effect of Char Loading. Another variable of interest is the effect of char loading (which dictates the gas residence time) on NO reduction and selectivity. The

Ind. Eng. Chem. Res., Vol. 42, No. 12, 2003 2541

Figure 10. Effect of the initial weight of char on the isothermal reduction of NO on bituminous coal char (flow, 1000 mL/min; inlet NO concentration, 1000 ppm; inlet oxygen concentration, 2.0%; temperature, 670 °C).

Figure 12. Trends in the concentration of NO and oxygen in a porous char undergoing heterogeneous gas-solid reactions.

the gas concentration. However, the gas concentration in a reacting porous solid is not constant throughout. Although the order of the reaction between carbon and NO and oxygen is a function of the reaction temperature and type of carbon, previous studies estimate the order to be unity in the temperature range employed in this study.24,25 Assuming that the carbon-oxygen and carbon-NO reactions are first, the differential equation, governing the instantaneous concentration of NO and oxygen in any porous particle, is reproduced as26 Figure 11. Effect of the initial weight of char on the selectivity of bituminous coal char for the char-NO reaction (flow, 1000 mL/ min; inlet NO concentration, 1000 ppm; inlet oxygen concentration, 2.0%; temperature, 670 °C).

1 d 2 dC D 2 r ) kaC dr r dr

(

)

(7)

with the boundary conditions being data reduction on an earlier set of experiments suggests that NO reduction can be enhanced by providing more carbon surface for the remaining NO molecules once oxygen is consumed. This can be achieved by increasing the initial weight of char in the reaction bed. A range of 500-1250 mg was tested, and the results of Figure 10 show that every increase in char loading leads to a higher extent of NO reduction. Further, the selectivity data, presented in Figure 11, indicate that every increase in char loading leads to an increase in NO selectivity. The reaction between NO and char is a type of gasification reaction and is thermodynamically favored to go to completion like the carbon-oxygen reaction. Hence, under any chosen process conditions, an increase in the carbon weight would lead to increased NO reduction, and even 100% NO reduction is possible, if the reaction is carried out under conditions that make NO the limiting reactant. Effect of the Particle Size. The last parametric study was carried out to quantify the effect of the initial particle size on the NO reduction extent and selectivity. As mentioned in the section on the temperature effect above, the char could become porous during the course of the reaction. The overall rate of any gas-solid reaction depends on the product of the reaction rate constant (k), a function of the reaction temperature and

r ) R; C ) Csurface

(8)

r ) 0; C is finite

(9)

and

The instantaneous solution to this differential equation is as follows:

(

C R sinhxka/Dr ) Csurface r sinhxka/DR

)

(10)

The radial concentration of gas through the solid depends on the ratio of the reaction rate constant of the particular gas-solid reaction and diffusivity (D) of the gas. If the reaction rate for the particular gas-solid reaction is high, the concentration of that gas falls off quickly in the radial direction toward the center of the particle and vice versa. The qualitative trends for the gas concentration as a function of radial position are shown in Figure 12 for the fast (carbon-oxygen) and slow (carbon-NO) reactions. It can be safely argued that the concentration of oxygen would fall more rapidly as we go away from the surface toward the particle’s center. Besides the porosity aspect, the particle size also

2542

Ind. Eng. Chem. Res., Vol. 42, No. 12, 2003

Figure 13. Simulation of the relative concentration of NO and oxygen in the radial direction assuming the ratio of reaction rate ) 20:1 in favor of oxygen: (4) NO concentration for 0.1 mm diameter; (2) O2 concentration for 0.1 mm diameter; (0) NO concentration for 1.0 mm diameter; (9) O2 concentration for 1.0 mm diameter; (O) NO concentration for 10 mm diameter; (b) O2 concentration for 10 mm diameter.

plays a very important role. If the particle is very small, then a significant drop in the concentration would not occur away from the surface and the concentration of the gas is significant throughout the entire particle. On the other hand, if the particle is very large, then the concentration drops off very quickly away from the surface. Although the experiments conducted in this study are not amenable to kinetic parameter estimation, a qualitative analysis can nevertheless be carried out. The molecular diffusivity of NO and oxygen can be estimated by the Wilke-Lee modification of the Hirschfelder-Bird-Spotz method.27 NO and oxygen are very similar in their molecular weight, size, etc. The estimated diffusivities of NO and oxygen in helium are 28.541 × 10-5 and 29.122 × 10-5 m2/s at 400 °C and 65.673 × 10-5 and 67.01 × 10-5 m2/s at 900 °C, respectively. Therefore, the radial concentration profile of each gas is a stronger function of the reaction rate constant. Arrhenius plots, based on the reactivity of cellulose char with a variety of gases (NO, oxygen, CO2, N2O, and H2O), estimate an order of magnitude ratio between the reaction rate constant for the carbon-NO and carbon-oxygen reactions.28 A simulation of the gas concentration, based on the estimated the diffusivities and reaction rate constants (with a ratio of 16 between carbon-oxygen and carbon-NO), is shown in Figure 13.21 It is possible to vary particle sizes so that the concentration of oxygen falls off rapidly from the surface while the NO penetrates significantly throughout the particle. In such a case, oxygen would be “filtered out” while a majority of the interior surface area would be “preferentially enriched” in NO compared to the bulk concentrations. Hence, even though the reaction rate constants for the two heterogeneous reactions remain constant, the overall rate of the carbon-NO reaction can be enhanced because of the alteration in the relative concentration of the gases in favor of NO. Such an enhancement is more effective in a 1.0 mm particle as compared to a 0.1 or 10 mm particle, as seen in Figure 13. The effect of the particle size, based on actual experimental data, is depicted in Figure 14. When the particle is relatively large (1.4-2 mm), the concentration of oxygen and NO both fall rapidly toward the center and only intermediate selectivity for the NO reduction is achieved. Similar NO reduction also occurs at a

Figure 14. Effect of the particle size on the isothermal reduction of NO on bituminous coal char (flow, 1000 mL/min; inlet NO concentration, 1000 ppm; wt of carbon, 1000 mg; inlet oxygen concentration, 2.0%; temperature, 670 °C).

Figure 15. Effect of the particle size on the selectivity of bituminous coal char for the char-NO reaction (flow, 1000 mL/ min; inlet NO concentration, 1000 ppm; wt of carbon, 1000 mg; inlet oxygen concentration, 2.0%; temperature, 670 °C).

relatively finer particle size range (250-500 µm) as well. The highest extent of NO reduction at 670 °C occurs for 710-850 µm. The selectivity trends corroborate the hypothesis. Figure 15 shows that the selectivity peaks at 90 mg of NO/g of char consumed for 710-850 µm particles. Conclusions The parametric study reveals that substantial reduction in NO is possible with the use of carbonaceous materials such as relatively inexpensive bituminous coal char. It is imperative for carbon-based processes to use inexpensive sorbents because they are consumed in this temperature range by the various gas-solid reactions. The feasibility of complete NO reduction from flue gas exists because the carbon-NO reaction, like the carbonoxygen reaction, is thermodynamically favored. The reaction can be carried out under a higher reaction temperature or a higher gas residence time such that the carbon-NO reaction is limited by NO and not

Ind. Eng. Chem. Res., Vol. 42, No. 12, 2003 2543

carbon. Whereas the temperature of operation is controlled by the availability of the flue gas at that temperature and the extent of CO emissions, the gas residence time, which controls the char loading, is limited by the pressure drop that can be afforded. Further, the bed height can be minimized by the use of optimized particle size char fractions, such as 710-850 µm identified under the process conditions employed in this study. It is evident that carbon reacts with oxygen to a higher extent. While this is not good for the selectivity of the carbon-NO reaction, the carbonoxygen reaction releases a significant amount of energy that can be harnessed for additional steam/electricity production in a power plant, thereby offsetting some of the operating costs associated with this carbon-based process. In fact, if all of the flue gas oxygen were consumed by carbon, this process would liberate a proportional amount of heat, reaching even 10% of the entire plant thermal output. Although the outlet oxygen concentration was not logged during these experiments, we had observed in another experiment that the exit oxygen concentration remained zero for over 30 min while the outlet NO concentration fell by only 50%.29 Data analysis reveals that both reactions (3) and (4) proceed to a similar extent. Therefore, the outlet CO emission under these experimental conditions would be about 1%. These observations are also supported by other reports that mention a higher exit CO/CO2 ratio with an increase in the reaction temperature for the carbon-NO reaction.30 In order for the process to be viable, it is imperative to identify techniques to reduce the reaction temperature to minimize CO emissions without incurring a loss in the extent of NO reduction. Acknowledgment The authors acknowledge Ohio Coal Development Office of the Ohio Department of Development for providing financial support for this project. Literature Cited (1) Muzio, L. J.; Quartucy, G. C. Implementing NOx Control: Research to Application. Prog. Energy Combust. Sci. 1997, 23, 233-266. (2) DOE. Reducing Emissions of Nitrogen Oxides via Low-NOx Burner Technologies; Clean Coal Technology Program, Topical Report No. 5; DOE: Washingotn, DC, Sept 1996. (3) Malone, P. Cardinal Unit 1. Large Scale Selective NonCatalytic Reduction Demonstration Project. Proceedings of the Sixteenth Annual International Pittsburgh Coal Conference: CoalEnergy and the Environment, Pittsburgh, PA, Sept 12-14, 1999. (4) DOE. Control of Nitrogen Oxide Emissions: Selective Catalytic Reduction; Clean Coal Technology Program, Topical Report No. 9; DOE: Washington, DC, July 1997. (5) Saxena, N.; Workosky, R. F.; Anderson, M. H.; Hwang, S.C. Removal of NOx and SOx Emissions from Pickling Lines for Metal Treatment. U.S. Patent 5,985,223, 1999. (6) DOE. Reburning Technologies for the Control of Nitrogen Oxide Emissions from Coal-Fired Boilers; Clean Coal Technology Program, Topical Report No. 14; DOE: Washington, DC, May 1999. (7) Burch, T. E.; Chen, W.; Lester, T. W.; Sterling, A. M. Interaction of Fuel Nitrogen with Nitric Oxide During Reburning with Coal. Combust. Flame 1994, 98, 391-401. (8) Chen, W.; Ma, L. Effect of Heterogeneous Mechanisms During Reburning of Nitric Oxide. AIChE J. 1996, 42 (7), 19681976.

(9) Kong, Y.; Cha, C. Y. NOx Adsorption on Char in the Presence of Oxygen and Moisture. Carbon 1996, 34 (8), 10271033. (10) Gangwal, S. K.; Howe, G. B.; Spivey, J. J.; Silveston, P. L.; Hudgins, R. R.; Metzinger, J. G. Low-Temperature Carbonbased Process for Flue Gas Cleanup. Environ. Prog. 1993, 12 (2), May, 128-132. (11) Knoblauch, K.; Richter, E.; Juntgen, H. Application of active coke in processes of SO2- and NOx- removal from flue gas. Fuel 1981, 60, Sept, 832-838. (12) Brown, C. E.; Hall, P. G. Physical Adsorption of Gases on Graphite. Trans. Faraday Soc. 1971, 67, 3558-3564. (13) Agnihotri, R.; Chauk, S. S.; Gupta, H.; Jadhav, R. A.; Mahuli, S. K.; Fan, L.-S. Reduction of NOx Emissions using Carbon and/or Carbon containing species. Proceedings of the Fifteenth Annual International Pittsburgh Coal Conference: CoalEnergy and the Environment, Pittsburgh, PA, Sept 14-18, 1998. (14) Teng, H.; Suuberg, E. M. Chemisorption of Nitric Oxide on Char. 2. Irreversible Carbon Oxide Formation. Ind. Eng. Chem. Res. 1993, 32, 416-423. (15) Smith, R. N.; Swinehar, N.; Lesnini, D. The Oxidation of Carbon by Nitric Oxide. J. Phys. Chem. 1959, 63, 544-547. (16) Mochida, I.; Ogaki, M.; Fujitsu, H.; Komatsubara, Y.; Ida, S. Reduction of nitric oxide with activated PAN fibers. Fuel 1985, 64, 1054-1057. (17) Furusawa, T.; Kunii, D.; Oguma, A.; Yamada, N. Rate of reduction of nitric oxide by char. Int. Chem. Eng. 1980, 20 (2), Apr, 239-244. (18) Aarna, I.; Suuberg, E. A review of the kinetics of the nitric oxide-carbon reaction. Fuel 1997, 76 (6), 475-491. (19) Chan, L. K.; Sarofim, A. F.; Beer, J. M. Kinetics of the NOCarbon Reaction at Fluidized Bed Combustor Conditions. Combust. Flame 1983, 52, 37-45. (20) Suzuki, T.; Kyotani, T.; Tomita, A. Study of the CarbonNitric Oxide Reaction in the Presence of Oxygen. Ind. Eng. Chem. Res. 1994, 33, 2840-2845. (21) DeGroot, W. F.; Richards, G. N. Gasification of Cellulosic Chars in Oxygen and in Nitrogen Oxides. Carbon 1991, 29 (2), 179-183. (22) Fan, L.-S.; Gupta, H. Method for the Treatment of Activated Carbonaceous Material containing Alkali/Alkaline Earth Metals for the Reduction of NOx from Flue Gas. U.S. Patent 6,224,839, 2001. (23) Gupta, H.; Agnihotri, R.; Jadhav, R. A.; Misro, S.; Fan, L.-S. The Influence of Oxygen on the Reduction of NOx Using Carbonaceous Materials. Proceedings of the Sixnteenth Annual International Pittsburgh Coal Conference: Coal-Energy and the Environment, Session 03, Pittsburgh, PA, Oct 12-14, 1999. (24) Li, Y. H.; Radovic, L. R.; Lu, G. Q.; Rudolph, V. A new kinetic model for the NO-carbon reaction. Chem. Eng. Sci. 1999, 54, 4125-4136. (25) Stanmore, B. R.; Brilhac, J. F.; Gilot, P. The oxidation of soot: a review of experiments, mechanisms and models. Carbon 2001, 39, 2247-2268. (26) Bird, R. B.; Stewart, W. E.; Lightfoot, E. N. Transport Phenomena; John Wiley and Sons: New York, 1960; pp 544-545. (27) Treybal, R. E. Mass-Transfer Operations, 3rd ed.; McGrawHill Book Company: New York, 1980. (28) Ruiz Machado, W. A.; Hall, P. J. Effects of Porosity on Carbon Reactivity in NO and O2. Energy Fuels 1998, 12, 958962. (29) Gupta, H.; Jadhav, R. A.; Misro, S.; Agnihotri, R.; Fan, L.-S. The Effect of Operating Prameters on the Selectivity of the Carbon-NO Reaction in an Oxidizing Atmosphere. Proceedings of the Seventeenth Annual International Pittsburgh Coal Conference: Coal-Energy and the Environment, Session 20, Pittsburgh, PA, Sept 12-14, 2000. (30) Teng, H.; Suuberg, E. M.; Calo, J. M. Studies on the Reduction of Nitric Oxide by Carbon: The NO-Carbon Gasification Reaction. Energy Fuels 1992, 6, 398-406.

Received for review September 10, 2002 Revised manuscript received February 4, 2003 Accepted February 5, 2003 IE020693N