Ind. Eng. Chem. Res. 1996, 35, 999-1005
999
Nitrogen Dioxide Reaction with Alkaline Solids Christopher H. Nelli and Gary T. Rochelle*,† Department of Chemical Engineering, The University of Texas at Austin, Austin, Texas 78712-1062
At conditions typical of a bag filter in a coal-fired flue gas, nitrogen dioxide (NO2) diluted with nitrogen was exposed to various alkaline and nonalkaline solids in a packed-bed reactor. On nonalkaline solids, 3 mol of NO2 reacted readily with water to produce 2 mol of nitric acid (HNO3) and 1 mol of NO. Alkaline solids such as calcium hydroxide or calcium silicate (from lime and fly ash) produced less NO and neutralized the nitric acid on the surface. A mathematical model developed to predict rates of NO2 removal on these surfaces successfully compared experimental and predicted rates. These results are relevant to technology for removal of NOx by addition of methanol to dry scrubbing systems for flue gas desulfurization. Introduction
Chemistry
The Clean Air Act of 1990 requires additional reduction of acid gases, sulfur dioxide, and nitrogen oxides released into the atmosphere from coal-fired electric power plants. The act is more flexible than previous air pollution legislation, allowing utilities to take a variety of routes to meet the more restrictive emissions requirements. In the case of older existing power plants, a possible route is to retrofit with a dry scrubbing process or make modifications to lime spray dryers already in place. A typical coal-fired power plant emits approximately 200-500 ppm of NOx (NO and NO2 combined), where the nitric oxide (NO) accounts for 90% of the NOx. Since NO is relatively unreactive and insoluble in aqueous solutions, a possible retrofit strategy is to oxidize the nitric oxide to nitrogen dioxide (NO2) by the addition of methanol or other hydrocarbons into the duct at an optimum temperature regime (Hori et al., 1992; Lyon et al., 1990). The focus of this research is to measure the reactivity of NO2 with various alkaline sorbents that potentially could be used to scrub NO2. In a typical dry scrubbing process, the solids are calcium based and high in surface area and porosity. The solids are free flowing but contain significant moisture due to their porosity. Once in the duct, the sorbents react with the flue gas and are removed by a particulate collection device such as a baghouse or electrostatic precipitator (ESP). Much of the reaction between the NO2 and the sorbent is expected to take place on a bag filter as moist, unreacted sorbent is contacted by oncoming flue gas. The relative humidity at the bag filter is expected to be 10-60% at a temperature of 60-90 °C. In this study, nitrogen dioxide (NO2) diluted with nitrogen was reacted with various alkaline solids in a packed-bed reactor. The primary reagent of interest was calcium silicate. Calcium silicate or ADVACATE (ADVAnced siliCATE) solids are comprised of varying amounts of calcium hydroxide reacted with fly ash (Kind and Rochelle, 1994). The reaction between silica in the fly ash and calcium hydroxide produces a calcium silicate material with high surface area and porosity. The other reagents of interest included hydrated lime (Ca(OH)2), calcium carbonate (CaCO3), and activated alumina.
The sorbent in the reactor at 50% relative humidity will be wet enough to assume a gas-liquid interface on the surface of the solids (either by multilayer coverage due to adsorption and deliquescence or by water-filled pores due to capillary action). The water on the surface allows absorbed gas species such as NO2 to be removed by pseudo aqueous phase reactions. Since NO2 and other nitrogen oxides play an important role in the production of nitric acid, there is an enormous amount of literature on their hydrolysis reactions and gas-phase equilibria. The equilibrium reactions in the gas phase are the following (Suchak and Joshi, 1994):
* Author to whom correspondence is addressed. † FAX: (512) 471-7060. E-mail:
[email protected].
0888-5885/96/2635-0999$12.00/0
2NO2(g) S N2O4(g)
(1)
NO(g) + NO2(g) S N2O3(g)
(2)
NO(g) + NO2(g) + H2O(g) S 2HNO2(g)
(3)
3NO2(g) + H2O(g) S 2HNO3(g) + NO(g)
(4)
The equilibrium constants for reactions (1-4) can be found in Suchak’s paper. Since only low concentrations of NO and NO2 were used in the experiments of this study (typically less than 400 ppm as opposed to 1-5% in the case of nitric acid production), equilibrium calculations showed no significant amounts of N2O4, N2O3, HNO2, and HNO3 to be produced in the gas phase. As a result, only NO2 absorbs into solution to react in the following manner (Suchak and Joshi, 1994):
NO2(g) S NO2(l)
(5)
2NO2(l) + H2O(l) S HNO2(l) + HNO3(l)
(6)
3HNO2(l) S HNO3(l) + 2NO(l) + H2O(l)
(7)
NO(l) S NO(g)
(8)
Reactions (5-8) can be added together to arrive at the following overall reaction:
3NO2(g) + H2O(l) S 2HNO3(l) + NO(g)
(9)
In other words, for every 3 mol of NO2 that absorbs and reacts, 1 mol of NO will be produced and desorb from solution, while 2 mol of HNO3 will accumulate in solution, neglecting loss due to its vapor pressure. Most researchers assume reaction (6) to be rate limiting and © 1996 American Chemical Society
1000
Ind. Eng. Chem. Res., Vol. 35, No. 4, 1996
Figure 1. Schematic of a packed-bed reactor system.
the other reactions to be in equilibrium (Counce and Perona, 1980). Takeuchi et al. (1977) and Shen and Rochelle (1995) have studied the absorption of NO2 into aqueous systems at low NO2 concentrations. For reaction (6), Takeuchi found that the hydrolysis of NO2 was secondorder with respect to NO2 and determined the second order rate constant at 25 °C to be 1.24 × 105 mol/L/ atm2/s. In addition, higher temperature lowered the rate of NO2 removal (an activation energy of -5.8 kcal/ mol was calculated from Takeuchi’s results). These results are consistent with the elementary steps that have been put forward regarding NO2 hydrolysis, i.e., 2 mol of NO2 dimerize to N2O4 before reacting with water. These steps allow one to reasonably conclude that the rate is second order in NO2 concentration and is faster at lower temperature since the NO2-N2O4 dimerization reaction favors N2O4 at lower temperature. Experimental Method The experiments were performed with a packed-bed reactor system (see Figure 1). A coarse glass frit (2 mm thickness) supported the solids within a cylindrical, Pyrex reactor (3.5 cm in diameter and 19.5 cm in height). Silica sand was used as a dispersant in order to prevent channeling and/or agglomeration of the reagents. A water bath with a PID controller regulated the temperature of the submerged reactor, while a syringe pump and furnace supplied humidity to the feed stream. The feed stream was synthesized by combining N2 with a commercially supplied gas containing 1% NO2 in N2. The reactor was equipped with a bypass to allow for preconditioning of the solids and to allow the gas concentration to stabilize before starting the experiment. Dry house air diluted the outlet stream of the reactor by a factor of 50 in order to reduce the NOx concentration to within the limits of the analyzer and also to reduce the humidity to prevent condensation downstream of the reactor. A portion of this stream was removed for analysis, while the balance was scrubbed with a 13 wt% NaOH solution. A Thermo Electron chemiluminescent NO/NO2/NOx analyzer (Model 14 B/E) was used to measure NO2 in the gas. For a typical experiment, sorbent mixed with sand was placed inside the reactor. The reactor contents
were preconditioned for 18 min with a nitrogen stream containing a known relative humidity. A flue gas bypassing the reactor was then synthesized with the same relative humidity as the preconditioning stream, allowed to reach steady state, and afterward sent into the reactor. The NO/NO2/NOx analyzer determined outlet concentrations from the reactor. Inlet concentrations into the reactor were measured when the system was in bypass mode at the beginning and end of the experiment. A gas-phase material balance around the reactor gave removal/production rates of NO2 and NO. To facilitate comparison of different solids, all of the experiments were conducted with a sufficiently high ratio of gas flow to alkali loading to avoid depletion of NO2 in the gas. Typically, the instantaneous NO2 removal from gas was less than 75% at short times and decreased to as little as 5% at long times. The primary reagent of interest in this study was calcium silicate or ADVACATE. In this study, the ADVACATE material contained 1 part Ca(OH)2 and 3 parts fly ash (Johnson, 1992). The details of the synthesis of this material can be found in Johnson’s thesis. Hydrated lime, produced by the Mississippi Lime Company, was the secondary reagent of interest. This material has been used by previous researchers for ADVACATE production (Kind and Rochelle, 1994; Arthur and Rochelle, 1995). The other reagents of this study, calcium carbonate and activated alumina, are reagent and chromatographic grade, respectively. The experiments were conducted at 25 and 70 °C. Relative humidity was varied from 0 to 80%. The feed stream into the reactor had NO2 concentration varying from 46 to 486 ppm at flow rates of 0.062-0.10 g‚mol/ min. Pressure was kept between 1.1 and 1.3 atm. Results The results from all of the experiments are tabulated and presented in Table 1. The experiments have been classified according to sorbent type: sand alone, hydrated lime, ADVACATE, calcium carbonate, ADVACATE produced from glass, activated alumina, and sand alone in the presence of 21% O2 in the gas stream.
Ind. Eng. Chem. Res., Vol. 35, No. 4, 1996 1001 Table 1. Experimental Results sand (g) sand
hydrated lime
ADVACATE
calcium carbonate
glass advacate alumina sand with O2
100 60 100 100 100 100 100 100 100 100 200 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 100
sorbent (g)
0.70 1.42 1.42 1.42 1.42 1.42 1.42 1.42 1.42 0.40 0.25 0.25 0.25 0.25 0.25 0.25 0.75 0.25 0.95 1.90 0.95 1.90 0.10 0.10 0.36
temp (°C)
RH (%)
pressure (atm)
flow (g‚mol/min)
NO2,in (ppm)
t ) 1 min
25 25 25 25 25 25 25 25 25 25 70 25 25 25 25 25 25 25 25 25 70 25 25 25 25 25 70 70 70 25 25 25 25 25 25 25 25
48 48 48 48 48 48 0 12 20 80 60 48 48 48 0 12 20 48 80 80 60 48 48 48 20 80 60 60 60 48 48 48 48 48 48 48 48
1.2 1.2 1.1 1.1 1.2 1.2 1.1 1.1 1.1 1.1 1.2 1.1 1.2 1.3 1.2 1.3 1.2 1.2 1.3 1.3 1.1 1.1 1.2 1.2 1.1 1.1 1.1 1.1 1.2 1.2 1.2 1.1 1.1 1.1 1.1 1.1 1.2
0.062 0.062 0.062 0.062 0.062 0.062 0.10 0.10 0.062 0.080 0.062 0.062 0.062 0.062 0.10 0.10 0.062 0.062 0.080 0.080 0.062 0.062 0.062 0.062 0.062 0.080 0.062 0.062 0.062 0.062 0.062 0.062 0.062 0.062 0.062 0.062 0.062
203 200 46 108 368 475 206 204 208 195 388 205 217 389 208 198 199 195 190 193 219 96 198 388 199 191 212 223 387 402 374 213 197 194 108 208 486
74 92 28 51 96 126 39 73 69 113 148 90 62 62 38 80 51 58 62 62 149 35 49 99 19 85 100 26 171 213 176 155 114 100 67 96 184
NO2,out (ppm) t)5 t ) 10 76 110 28 54 172 222 139 123 75 135 194 88 62 52 33 84 52 55 62 63 158 47 71 119 48 108 132 75 205 245 191 165 113 105 69 105 189
t ) 20
114 154 29 59 296 342 199 180 126 135 297 91 55 45 36 82 49 55 62 63 161 52 78 139 68 115 140 91 220 256 194 166 114 110 72 109 282
166 177 30 71 341 419 204 193 192 139 354 100 60 61 64 83 45 53 66 166 56 89 169 82 122 146 103 241 275 217 170 122 116 75 120 385
reaction can be expressed in terms of partial pressures and derived as the following:
[
2
rate ) kbulk PNO2 -
Figure 2. Raw data of a typical experiment with sand alone in the reactor.
Mathematical Model for Sand Systems To gauge the effect of sand, NO2 removal was measured when sand alone was in the reactor. The presence of sand caused significant NO2 removal along with NO production. Raw data from a typical experiment is shown in Figure 2. At all times, the sand produced a stoichiometric ratio of 1 mol of NO produced per 3 mol of NO2 removed. This surprising amount of removal was assigned to the mechanism as shown in reactions (5-9). A full attempt in modeling the observed phenomena was undertaken. The following mathematical model is based upon the removal mechanism as presented in reactions (5-9). If reaction (6) is the rate-limiting step and vapor-liquid equilibrium is assumed, the rate of
]
PHNO34/3PNO2/3 PH2O2/3K42/3
)
kbulk[PNO22 - PNO22*] (10)
where kbulk is the forward rate constant for reaction (6), Px is the partial pressure of x, Px* is the equilibrium partial pressure of x, and K4 is the gas-phase equilibrium constant for reaction (4). Even though the reactions behind this derivation are seen in bulk liquid systems, the relationship between the forward and reverse rates in eq 10 will be assumed true in any environment. However, the form of the rate constant, kbulk, will change depending on the nature of the reaction, i.e., surface or bulk liquid. To compare model results with experimental results, material balances were done on HNO3 in the pseudo liquid phase on the surface of the sand and on NO2 in the gas phase, respectively:
[
]
PHNO34/3PNO2/3 d[HNO3] 2 ) kbulk (CNO2RT) dt P 2/3K 2/3 v
H2O
dCNO2 dx
) -(3/2)
4
[]
d[HNO3] Vl dt Vr
(11)
(12)
where v is the gas velocity, Cx is the gas concentration of x, [HNO3] is the liquid concentration of HNO3, Vl is
1002
Ind. Eng. Chem. Res., Vol. 35, No. 4, 1996
Figure 3. NO2 removal by sand alone. Curves predicted by eqs 10-14. Rate constant (kbulk) ) 2.4 × 105 mol/L/atm2/s. BET surface area of sand ) 0.57 m2/g.
Figure 4. NO2 removal by sand alone as a function of HNO3 accumulation. Curves and HNO3 calculated by eqs 10-14. Rate constant (kbulk) ) 2.4 × 105 mol/L/atm2/s.
the volume of water on the surface of the sand, and Vr is the volume of the reactor. The following equations define the pressure of HNO3 and NO over solution:
PHNO3 ) f([HNO3]) PNO ) (1/2)
d[HNO3] VlRT/G dt
(13) (14)
where G is the gas flow rate and f([HNO3]) is a polynomial fit of the vapor-liquid equilibrium data gathered by Davis and de Bruin (1964). The amount of water on the surface of the sand was calculated by assuming a water molecule has six nearest neighbors, the diameter is equal to its collision diameter, and the surface area and number of monolayers is best approximated by BET theory. In the NO2 balance, the partial derivative with respect to time has been neglected. This approximation is valid if small step sizes are taken when stepping through time during numerical integration. Euler’s method was the numerical method chosen to solve the coupled ordinary differential equations. This model has one adjustable parameter, the reaction rate constant kbulk. The initial conditions for the HNO3 and NO2 material balances were their respective concentrations at t ) 0 ([HNO3] ) 0) and x ) 0 (CNO2 ) inlet concentration). The rate of NO2 removal was calculated by multiplying the difference in inlet and outlet concentration with the gas flow rate, G. The results of the modeling can be seen in Figures 3 and 4, where sand amount and NO2 concentration were varied, respectively. A good fit was obtained between the experimental and predicted rates. The rate is greatest initially since the solution on the surface is clean of HNO3. However, as HNO3 accumulates on the surface, the rate slows as reaction (6) approaches equilibrium. Finally, at long times, the removal rate asymptotically approaches zero. The model consistently predicted lower rates of removal in this region. However, it should be noted that experimental precision was probably poor at long times since removal was less than 10% of the inlet concentration. An overall rate constant of 2.4 × 105 mol/L/atm2/s yielded the best match between experimental and predicted values. To further test the model, the relative humidity was varied from 0 to 80%. As seen in Figure 5, the rate of NO2 removal was inversely proportional to the relative humidity of the inlet gas stream. The highest rate occurred when a feed stream with no moisture was fed into the reactor (in reality, residual H2O remained on
Figure 5. Effect of relative humidity on NO2 removal by silica sand. Curves calculated by eqs 10-14 with a different rate constant, kbulk (mol/L/atm2/s), for each relative humidity.
the surface since NO was still produced at a 1 to 3 mole ratio to NO2 removed). Increasing the amount of water on the surface decreased the rate, but the capacity to sustain removal was increased. In order to match experimental and predicted values, the rate constant kbulk was varied. The experiment with 0% relative humidity was not modeled due to the difficulty in estimating Vl. The negative influence of Vl on the rate constant strongly indicates the reaction is surface catalyzed. When more water than necessary was on the surface, the water seems to foul the surface reaction. Since the amount of NO produced per NO2 removed remained constant at a mole ratio of 1 to 3, the negative effect of relative humidity on the rate constant suggests water coverage of sites on the surface of the sand rather than a change in mechanism. Mathematical Model for Alkaline Systems The following mathematical model for alkaline systems will assume the rate of NO2 removal is second order in NO2 concentration for all surfaces. Furthermore, the rate of HNO3 accumulation will be assumed to be zero since the calcium-based sorbents are initially expected to completely neutralize the HNO3 produced on the surface. However, for nonalkaline substances like sand, the model will apply only at early times while HNO3 concentration is still low. Once again, the model consists of material balances on HNO3 in the liquid phase and NO2 in the gas phase:
Ind. Eng. Chem. Res., Vol. 35, No. 4, 1996 1003
d[HNO3] )0 dt v
dCNO2 dx
(15)
[ ] [ ] Vlsand
) -(3/2)kbulk,sand[CNO2RT]2
Vr
Vlalk
(3/2)kbulk,alk[CNO2RT]2
Vr
-
(16)
where Vlx is the volume of water on material x and kbulk,x is the rate constant on surface x. The model was first applied to the sand only system in order to evaluate kbulk,sand. After which, calciumbased alkaline sorbents and activated alumina were added to the sand and tested. The sole parameter kbulk,alk was adjusted to give the best match between experimental and predicted rates. This comparison of rates is shown in Figure 6 for the hydrated lime-sand system. Both sorbent amount and NO2 concentration were varied. The results verify the assumption that the rate is second order in NO2 concentration. Similar graphs were obtained for the other sorbents. Table 2 presents the rate constants and BET surface areas for the sorbents tested. Two forms of the rate constant are shown in Table 2. These constants are specific to where the reaction is assumed to be occurring, such as in the bulk liquid or on the surface. For the surface values, the following equation was used:
ksurface ) kbulk
[
]
Vl surface area
(17)
Figure 6. NO2 removal by hydrated lime. Initial rates predicted by eqs 15 and 16. Rate constant on sand (kbulk,sand) ) 2.4 × 105 mol/L/atm2/s. Rate constant on hydrated lime (kbulk,alk) ) 1.1 × 106 mol/L/atm2/s. Table 2. Comparison of Rate Constants (25 °C and 48% RH) kbulk × 10-5 ksurface surface area (mol/L/atm2/s) (mol/m2/atm2/s) (m2/g) sand calcium carbonate hydrated lime ADVACATE alumina Takeuchi et al. (1977)a b
2.4 2.4 11 16 0.90 0.62b
0.077 0.077 0.34 0.51 0.029
0.57 5.47 8.76 49.9 207
a Investigated NO -H O reaction with a stirred-cell reactor. 2 2 Rate constant modified to account for stoichiometry.
A relatively small rate constant was found for the alumina sorbent. Inadequate diffusion into this highly porous material lowered the observable rate constant below its true rate constant. This effect is studied in greater detail in a later section of this paper. Effect of Sorbent Type on NO Production Figure 7 shows the ratio of NO produced per NO2 removed for various substances. As mentioned before, sand alone in the reactor produces a ratio of 1 to 3. Oxygen was added to the feed stream in anticipation of oxidizing the NO to NO2 and thus achieving ratios less than 0.33. However, under the conditions of low NO concentration and pressure, the presence of 21% O2 in the feed stream did not alter the ratio or otherwise affect the rate of NO2 removal. The other three experiments involved alkaline sorbents mixed with sand. The alkaline sorbents produced ratios less than 0.33. Initially, the alkalinity provides a high enough pH to keep HNO2 and HNO3 dissociated. As a result, HNO2 is unable to decompose into NO (reaction (7)). At some point, the solution on the surface begins to precipitate Ca(NO2)2 and Ca(NO3)2. Due to product formation, the alkalinity is reduced since it is more difficult for the hydroxide to reach the acid on the surface. As pH decreases, HNO2 no longer dissociates and NO production starts. Eventually, none of HNO2 dissociates and the sorbent reaches a final ratio of 0.33. Instantaneous removal, however, is still higher than the sand alone case since HNO3 is still being removed from the surface, albeit at a much slower rate than before. The sorbent is spent when the hydroxide can no longer reach the surface to control the HNO3 concentration. It should be noted that, for the sorbent experiments, different mass amounts were used to achieve a common sorbent surface area. Thus, the ADVACATE material, having the least amount of mass and calcium as well
Figure 7. Sorbent effect on NO production at 25 °C and 48% relative humidity. Loading for alkaline experiments was 12.5 m2 sorbent and 11.4 m2 sand. Conversion assumes stoichiometry of 0.5 mol of Ca2+/mol of NOx removed.
as the highest rate constant, achieves the highest conversion and a ratio that is closest to 0.33. The calcium carbonate, having the lowest surface area and rate constant, achieves the lowest conversion. The downward trend in the calcium carbonate case is due to NO production on the sand that is greatest initially but is quickly reduced when the sand shuts down due to HNO3 accumulation. This effect is only seen with the calcium carbonate, since its kbulk,alk is comparable to kbulk,sand. Effect of Particle Size, Relative Humidity, and Temperature Figure 8 shows the effect of particle size on the rate of NO2 removal. Two types of ADVACATE material were used: one that used fly ash as a source of silica,
1004
Ind. Eng. Chem. Res., Vol. 35, No. 4, 1996 Table 3. Effect of Temperature and RH on ksurface (mol/m2/atm2/s) 20% RH, 25 °C
48% RH, 25 °C
80% RH, 25 °C
60% RH, 70 °C
2.0 0.54 0.12
0.51 0.34 0.077 0.077 0.029
0.33 0.45 0.029
0.22 0.23 0.018
ADVACATE hydrated lime sand calcium carbonate alumina
Conclusions
Figure 8. Effect of particle size on NO2 rate of removal as a function of NO2 inlet concentration.
the other that used recycled glass as a source of silica. The starting material for the two types of ADVACATE had different diameters (approximately 5-10 µm for the fly ash and approximately 25 µm for the glass). During typical ADVACATE production, the particle size of the final material is the same size as the starting material. Thus, the fly ash ADVACATE is assumed to have a lower particle diameter than the recycled glass ADVACATE. However, both materials were assumed to have the same reactivity toward NO2. Preliminary calculations suggested that inadequate diffusion in the larger glass ADVACATE particle would reduce NO2 removal more than slight changes in reactivity. In order for a material to have a large surface area (such as ADVACATE), most of the surface area must be found internally in pores. If a surface reaction is involved, the larger particle requires a deeper NO2 penetration in order to be as effective as the smaller particle. Otherwise, the surface area at the center of the larger particle is not being utilized as equally as the center of the smaller particle. This effect is shown in Figure 8, where the larger particle size material produced the lower rate at inlet NO2 concentrations of 100 and 200 ppm. The concentration profile within the particle is a function of both the diffusion rate and reaction rate. For a second-order reaction, the concentration profile within the particle becomes more diffusion controlled at lower concentrations since the reaction rate is reduced more than the diffusion rate. This will tend to equalize the relative penetration depth of NO2 between the two different size particles. The results in Figure 8 show the validity of this hypothesis. The rates of the two different particle sizes at the lower concentration are closer together than the rates at the higher concentration. This, in itself, supports the assumption that the reaction is second order in NO2 concentration. For if the reaction was first order in NO2 concentration, then the relative difference of the rates would not change as inlet NO2 concentration was varied. For the alkaline materials, the effect of relative humidity on the NO2 rate of removal was similar to the sand alone case (see Table 3). A general decrease in the surface rate constant was observed for all the materials tested. As expected from the literature (Takeuchi et al., 1977), the rate of NO2 removal decreased with rising temperature (see Table 3). This effect is the result of an intermediate step in the reaction mechanism. The step, in which NO2 dimerizes to N2O4, is less favorable at higher temperatures.
Sand alone in the reactor removed NO2 and produced NO. A full and complete mathematical model was developed for the sand alone system with only one adjustable parameter. A rate constant of 2.4 × 105 mol/ L/atm2/s yielded the best match between experimental and predicted values at 48% relative humidity. The rate constant was adjusted in response to changes in relative humidity. As a result, the reactions are seen to be surface catalyzed. For systems involving alkaline sorbents, a simpler and less encompassing model was adopted. This one-parameter model was able to measure the reactivity of various materials toward NO2. The addition of alkaline materials to the system resulted in less NO production. Finally, increases in the sorbent particle size, relative humidity, and temperature decreased the rate of NO2 removal. Acknowledgment This work was supported by a grant from the Texas Advanced Technology Program (TATP). Nomenclature Px ) partial pressure of x, atm K4 ) gas-phase equilibrium constant for reaction (4), 0.0102 1/atm v ) gas velocity, cm/s x ) length of reactor, cm Cx ) gas concentration of x, mol/L R ) gas constant, L atm/mol/K T ) temperature, K Vl ) liquid volume of water on the surface of sand, L Vr ) gas volume of reactor, L t ) time, s [HNO3] ) liquid concentration of HNO3, mol/L rate ) rate of reaction (6), mol/L/s G ) gas flow rate, L/s ksurface ) surface reaction rate constant, mol/m2/atm2/s kbulk ) bulk liquid reaction rate constant, mol/L/atm2/s M ) molarity min ) minute RH ) relative humidity
Literature Cited Arthur, L. F.; Rochelle, G. T. SO2 Removal by Reagents Prepared from Lime and Recycled Glass. Presented at the SO2 Control Symposium, Miami, FL, 1995. Counce, R. M.; Perona, J. J. A Mathematical Model for Nitrogen Oxide Absorption in a Sieve-Plate Column. Ind. Eng. Chem. Process Des. Dev. 1980, 19, 426. Davis, W., Jr.; de Bruin, H. J. New Activity Coefficients of 0-100 Percent Aqueous Nitric Acid. J. Inorg. Nucl. Chem. 1964, 26, 1069. Hori, M.; Matsunaga, N.; Malte, P. C.; Marinov, N. M. The Effect of Low-Concentration Fuels on the Conversion of Nitric Oxide to Nitrogen Dioxide. Presented at the Twenty-Fourth Symposium (International) on Combustion/The Combustion Institute, Sydney, Australia, 1992, p 909.
Ind. Eng. Chem. Res., Vol. 35, No. 4, 1996 1005 Johnson, H. L. The Effect of Moisture on the Reaction of Sulfur Dioxide with Calcium Silicate Sorbents. M.S. Thesis, Department of Chemical Engineering, The University of Texas at Austin, Austin, TX, 1992. Kind, K. K.; Rochelle, G. T. Preparation of Calcium Silicate Reagent from Fly Ash and Lime in a Flow Reactor. J. Air Waste Manage. Assoc. 1994, 44, 869. Lyon, R. K.; Cole, J. A.; Kramlich, J. C.; Chen, S. L. The Selective Reduction of SO3 to SO2 and the Oxidation of NO to NO2 by Methanol. Combust. Flame 1990, 81, 30. Shen, C. H.; Rochelle, G. T. NO2 Absorption in Limestone Slurry for Flue Gas Desulfurization. Presented at the SO2 Control Symposium, Miami, FL, 1995. Suchak, N. J.; Joshi, J. B. Simulation and Optimization of NOx Absorption System in Nitric Acid Manufacture. AIChE J. 1994, 40 (6), 944.
Takeuchi, H.; Ando, M.; Kizawa, N. Absorption of Nitrogen Oxides in Aqueous Sodium Sulfite and Bisulfite Solutions. Ind. Eng. Chem. Process Des. Dev. 1977, 16, 303.
Received for review February 16, 1995 Accepted June 9, 1995X IE950117+
X Abstract published in Advance ACS Abstracts, February 15, 1996.
