Integrated System for the Treatment of Oxides of Nitrogen from Flue

Department of Civil Engineering, Indian Institute of Technology, Madras, ..... To study the effect of sulfur dioxide on oxidation by ozone, the experi...
0 downloads 0 Views 312KB Size
Environ. Sci. Technol. 2006, 40, 1035-1041

Integrated System for the Treatment of Oxides of Nitrogen from Flue Gases SANJOY BARMAN AND LIGY PHILIP* Department of Civil Engineering, Indian Institute of Technology, Madras, India 600 036

A novel and effective system was developed for the complete treatment of NOx from flue gases. The system consisted of photocatalytic or ozone oxidation of NOx, followed by scrubbing and biological denitrification. Maximum photocatalytic oxidation of NOx was achieved while using powdered TiO2 at a catalytic loading rate of 10 g/ h, relative humidity of 50%, and a space time of 10 s. The used catalyst was regenerated and reused. A total of 72% of oxidized NO was recovered as HNO3/HNO2 in the regeneration process. Stoichiometrically, 10% excess ozone was able to affect 100% oxidation of NO to NO2. Presence of SO2 adversely influenced the oxidation of NO by ozone. The scrubbing of NO was effective with distilled water. Heterotrophic denitrifiers were able to denitrify the leachate with an efficiency of 90%, using sewage (COD 450 mg/L) as electron donor. The new integrated treatment system seems to be a promising alternative for complete treatment of NOx from flue gases.

Introduction In general, oxides of nitrogen (NOx) refer collectively to six compounds of nitrogen and oxygen. However, it often refers to the major species: nitric oxide (NO) and nitrogen dioxide (NO2). NOx is responsible for troposphere ozone and urban smog through photochemical reactions. NOx, together with SO2, is the major contributor to acid rain that harms forest crops, buildings, as well as aquatic life (1, 2). The rapid economic growth and ever increasing consumption of fossil fuels have resulted in large emissions of NOx (3, 4). Concern for environmental and health issues have forced the environmental regulatory agencies to enforce stringent NOx emission standards. For example, the 1990 Clean Air Act Amendments have led to regulations requiring significant NOx emission reduction from stationary sources in the U.S. (5). The recently promulgated Clean Air Interstate Rule (CAIR) enforces a large reduction in NOx emissions. CAIR plans for 60% NOx and 70% SO2 reduction from 2003 levels within the next 10 years (6). Various methods exist to reduce NOx emission. Combustion modification and selective catalytic reduction (SCR) methods are probably the most widely used techniques to control NOx emissions from industries (7-9). However, reduction in NOx is often limited in the combustion modification methods, while SCR systems can be expensive. Other new technologies such as nonthermal plasma and pressure swing adsorption appear to be efficient and costeffective for the removal of higher concentrations of NOx, * Corresponding author phone: +91-44-257-4274; fax: +91-44257-4252; e-mail: [email protected]. 10.1021/es0515102 CCC: $33.50 Published on Web 12/21/2005

 2006 American Chemical Society

but they are still expensive for the treatment of huge volumes of flue gases (10, 11). Thus, there is a need for environmentally friendly and cost-effective alternatives for comprehensive treatment of NOx from flue gases. Biological removal of NOx from contaminated gas stream is emerging as a novel treatment method. Biofiltration, or the use of microorganisms to treat air streams, seems to be a more promising alternative to conventional air pollution control technologies (12, 13). It has been reported that NO2 and SO2 can be removed effectively using a biotrickling filter/ scrubber within a contact time of 6 s, due to their high solubility in water (14, 15). However, it may be noted here that NO represents 85-95% of the total NOx generated in the combustion process. A few attempts have been made in the past to find efficient biological methods for NOx removal from flue gases (16-18). They employed denitrification or nitrification processes. Davido et al. (18) demonstrated the potential of nitrifying bacteria for the removal of NO. The system required a long residence time of 13.7 min to remove 90% NO from a 100 ppm contaminated stream. Up to 96% removal of NO was observed with Thiobacillus denitrificans for a gas stream containing 5000 ppm of NO. However, simultaneous SO2/NOx removal from flue gas was not technically feasible by a combined system with Desulfovibrio desulfuricans and T. denitrificans due to the NO inhibition to D. desulfuricans (19, 20). The presence of 3-8% oxygen in the flue gases of utility boilers adversely affected the anoxic denitrification process, whereas the nitrification process needed a very high empty bed contact time (EBRT) on the order of 10-14 min. Due to the large volume of NO generated and its very low solubility (62 mg in 1 kg of water at 20 °C and an NO pressure of 760 mmHg) (21), neither biotrickling filters nor scrubbers seem to be a viable option for the removal of NO within practical contact times. Accordingly, conversion of NO to NO2, or any other soluble form using a suitable technique, followed by scrubbing and denitrification seems to be a viable research area that could lead to a cost-effective and practical NOx reduction alternative. Titanium dioxide is gaining importance as a photocatalyst for treatment of a wide range of organic and inorganic pollutants (22-25). It has been reported that NOx can be effectively oxidized to the soluble form/forms by means of photocatalytic oxidation using TiO2 (24-26). However, optimization of the operating parameters and assessment of the maximum conversion efficiency of this approach need to be addressed. Chemical oxidation is another promising process which can effectively oxidize most of the compounds. Ozone (O3) is one of the most reactive oxidizing agents. The feasibility and plausible mechanisms of this process were demonstrated by Puri (27). In the present study, development of a new integrated system for the complete treatment of NOx from flue gases is described. The treatment system consisted of advanced oxidation (photocatalytic oxidation using TiO2 and oxidation by ozone) followed by a scrubber and/or a denitrification process. The proof of concept was demonstrated at the benchscale. The performance of the system was monitored under selected operating conditions. An attempt was also made to recover the reusable byproducts of the process.

Materials and Methods Photocatalytic Reactor. A schematic of the reactor is shown in Figure 1a. The rectangular shaped reactor (dimension: 50 cm × 12 cm × 12 cm) was made of acrylic sheets of 6 mm thickness. Baffles (12 cm × 8 cm × 0.6 cm) were provided on both sides of the reactor at 5 cm spacing to get adequate VOL. 40, NO. 3, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1035

FIGURE 1. (a) Schematic of the photocatalytic reactor; (b) schematic of the scrubber; (c) process flow diagram for the complete treatment of NOx. contact between catalyst and NOx. A 15 W UV light was fitted on either side of the reactor to irradiate the TiO2 coated baffles. The entire set up was covered with a lid made of aluminum sheet. The reactor was operated at 27 °C and atmospheric pressure, with a space time (ST) of 10 s. Air with a relative humidity (RH) of 80% was mixed with NOx to obtain the desired concentration. NOx was supplied from an NO cylinder (5125 ppm) (Indo Gas, Chennai, India). Flow rates of both NOx and air were maintained by a rotameter. Inlet NOx concentration was maintained at 300 ppm (NO, 260 ppm; NO2, 40 ppm) and ST (10 s) was adjusted by controlling the streamflow rate. TiO2 (20 gm) was first mixed with distilled water to make a paste. The baffles were first painted with gum (Fevicol) and then with TiO2 slurry over the gum layer. They were then dried by keeping it in an oven at 105 °C for 24 h. The catalyst loading in the reactor was 1.5 mg/cm2. The reactor was also operated, with TiO2 in powder form, by varying the relative humidity (RH 10-80%) and catalyst loading rate (4-10 g/h). 1036

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 3, 2006

Oxidation by Ozone. Oxidation of NOx by ozone was performed in a 5 L Borosil bottle at room temperature (27 °C) and atmospheric pressure. Air (RH 80%) and NOx streams were mixed to obtain the desired concentration of NOx. Inlet NOx concentration was maintained at 200 ppm (NO, 175 ppm; NO2, 25 ppm) and ST (10 s) was adjusted by controlling the inlet gas flow rate. Ozone was supplied from an ozone generator (Vortex, India), and flow was controlled with a flow controller. Gas stream and ozone entered the reactor through different ports. NO and NO2 concentrations in inlet and outlet were measured using a combustion gas analyzer (Quintox KM 9106, Kane May, U.K.) at 5 min intervals. The same experiment was also performed in the presence of SO2 (100 ppm) in the inlet gas stream. Scrubber. Nitrogen dioxide from the ozone oxidizing reactor (Figure 1b) was scrubbed at room temperature (27 °C) and atmospheric pressure. The reactor was made of clear acrylic pipe. The total length of the reactor was 75 cm with a diameter of 6 cm and a bed depth of 60 cm. The reactor

TABLE 1. Characteristics of Sewage sample

parameter

value

1 2 3 4 5 6 7 8 9 10 11 12 13 14

pH dissolved solids (mg/L) total solids (mg/L) volatile solids (mg/L) fixed solids (mg/L) dissolved oxygen (mg/L) total COD (mg/L) soluble COD (mg/L) BOD5 (mg/L) alkalinity (mg/L as CaCO3) chlorides(mg/L) sulfate (mg/L) iron (mg/L) copper (mg/L)

7.4 280 920 300 620 0.0 824 450 180-230 405 631 125 2.5 1.5

contained 1.69 L of packing made of plastic rings. Void fraction of the packing was measured as 0.75. The trickling liquid was sprinkled over the packed bed from the top. The trickling rate was varied between 1 and 10 L/h. The trickling filter effluent was collected from the bottom of the reactor and fed to the biological denitrification system. The gas inlets and outlet ports were located at the bottom and top lids of the reactor, respectively. The EBRT was varied from 3 to 8 s. For each residence time and trickling liquid flow rate, the scrubbing efficiency was monitored. NO and NO2 concentrations were measured at inlet and outlet ports of the reactor. Scrubbed liquid was analyzed for nitrite (NO2-) and nitrate NO3-. Distilled water and mineral media (CaCl2‚2H2O, 50 mg; MgSO4‚7H2O, 216 mg; K2HPO4, 1 g; KH2PO4, 1 g; in 1 L distilled water) with and without sodium bicarbonate (1 g/L) were used as trickling/scrubbing liquids. Process Flow Diagram. The flow diagram of the processes used for the complete treatment of NOx is given in Figure 1c. The process flow diagram has two alternatives represented by solid and dashed lines. In the first alternative, NOx treatment is carried out by photocatalytic oxidation followed by oxidation by ozone, scrubbing, and biological denitrification, whereas in the second alternative, it was achieved by oxidation of NOx by ozone followed by scrubbing and biological denitrification. The spent TiO2 was regenerated using distilled water. The spent TiO2 powder (5 g) was collected in a reaction bottle containing 250 mL of distilled water and was kept in a shaker at 150 rpm for 2 h. The solidliquid separation was achieved by centrifuging the reaction mixture. The regenerated TiO2 was dried and reused. Denitrification. Heterotrophic denitrifying bacterial consortium was developed from the seed activated sludge collected from Koyambedu Sewage Treatment Plant, Chennai, India, by enrichment technique. The seed sludge was initially fed with 200 mg/L of nitrate and ethanol as electron donor. Once the nitrate removal of the system was above 60%, nitrate concentration in the feed was gradually increased up to 1000 mg/L. The composition of nutrient medium used for the study was as follows: CaCl2‚2H2O, 50 mg; MgSO47‚ H2O, 216 mg; NaNO3, 20 mM; ethanol, 10 mM; K2HPO4, 1 g; and KH2PO4, 1 g in 1 L distilled water. Kinetic studies were conducted using both ethanol and sterilized sewage as electron donors, and the data was used for biokinetic parameter estimation. The characteristics of sewage used in the present study are given in Table 1. These investigations were carried out in 500 mL reaction bottles (Borosil, India) with a nylon septum. The reaction mixture consisted of 50 mL of denitrifying bacterial culture (3000 mg/L) and 250 mL of nutrient media prepared in sterilized sewage (COD ) 450 mg/L), spiked with desired nitrate concentration. To find out the minimum COD:N ratio for effective denitrification, kinetics of denitrification was also carried out by varying the

FIGURE 2. Conversion of NO to NO3- by photocatalytic oxidation using TiO2 as catalyst (space time, 10 s; inlet NOx concentration, 300 ppmv; relative humidity, 80%). nitrate concentration from 200 to 2000 mg/L, keeping the electron donor concentration (COD) as a constant. Denitrification of scrubber leachate was carried out in a sequential batch reactor of 6 L capacity and a biological sludge retention time (BSRT) of 6 days. The biomass concentration in the reactor was 3000 mg/L (as MLSS) with a reaction time of 6 h, a settling time of 1 h, and fill and drain times of 30 min each. Analyses. The analyses of NOx and SO2 were carried out using a combustion gas analyzer (Quintox KM 9106, Kane May, U.K.). Selected grab samples were analyzed using NOx and SO2 Draeger tubes (Fisher Scientific, U.S.A.). The lower detection limit of each method was 1 ppmv. Nitrate/nitrite analysis was carried out as per standard methods (28). COD analysis was conducted by closed reflux method using a HACH COD digester (HACH, U.S.A.) as per standard methods. pH measurements were carried out using a regular pH meter (Cyber Scan 510, India).

Results and Discussion Photocatalytic Oxidation of NOx. The photocatalytic oxidation of NO was carried out in a reactor with TiO2 coated baffles having a catalyst loading of 1.5 mg/cm2. Conversion of NO was relatively high (46% in 30 min) in the beginning. It then decreased gradually and approached a steady state (20% in 150 min) as shown in Figure 2. It is well-known that TiO2 irradiated with light shorter than 380 nm wavelength generates positive holes (h+) and photoelectrons (e-), which catalyze both oxidation and reduction reactions. The photo catalytic oxidation of NO on the surface of TiO2, in the presence of oxygen and water vapor, is well explained elsewhere (25). As a result of this reaction, nitrite and nitric acid were generated. The initial conversion might be due to the high initial rate of adsorption plus reaction (chemisorption) of NO. This might be because initially almost all the catalyst surface was available for adsorption of NO. As time progressed, more and more nitrate and other oxidized species might have deposited on the catalyst, blocking available surfaces, which in turn, decreased the oxidation efficiency. Photocatalytic Oxidation with TiO2 in Powder Form. As discussed above, conversion of NOx was not adequate while using TiO2 coated baffles inside the reactor. Since photocatalytic oxidation is a heterogeneous reaction, its conversion efficiency depends on specific surface area (area per unit mass) and the contact between the pollutant and the catalyst (25). To get a high specific surface area and intimate contact, experiments were conducted by introducing TiO2 powder into the inlet gas stream, with uncoated baffles in the reactor. (a) Effect of TiO2 Concentration. Different amounts of TiO2 were introduced into the reactor, keeping the same residence time, inlet NOx concentration, and light source as in the previous case. It was found that conversion increased VOL. 40, NO. 3, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1037

FIGURE 3. Conversion of NOx with powder TiO2 (space time, 10 s; inlet NOx concentration, 300 ppmv): (a) effect of catalytic loading rate (RH, 80%); (b) effect of relative humidity (catalytic loading, 10 g/h). with an increase in amount of TiO2. Around 60% conversion with a TiO2 loading rate of 10 g/h was noted, whereas it was only 35% when the catalyst loading rate was 4 g/h (Figure 3a). It is obvious that the total surface area depends on the amount of material. Relatively higher conversion of NOx with a higher catalyst loading rate can be attributed to the larger surface area available for the reaction. (b) Effect of Humidity. To study the effect of humidity on the photocatalytic oxidation of NOx, the experiment was conducted with varying humidity. Relative humidities of 10, 50, and 80% were maintained in the inlet stream, with a TiO2 loading rate of 10 g/h. Conversion of NO increased with an increase in humidity up to 60% (Figure 3b). The efficiency of NOx removal from the air stream by TiO2 was affected by the relative humidity of the test environment. Devahasdin et al. (24) reported that the NO conversion efficiency increased with respect to the increase in humidity and became asymptotic at a humidity of 50%. During photocatalysis, the continuous consumption of hydroxyl radicals requires replenishment to maintain catalyst activity. With an increase in moisture content, water adsorption is favored, which in turn enhances electron-hole recombination and increased conversion. At high humidity (80%), a marginal decrease in conversion (around 60%) against 50% RH (conversion: 65%) was observed. This may be due to ready agglomeration of powdered TiO2 at higher humidity, which decreases the available surface area. Regeneration of Used Catalyst. For the effectiveness of any catalytic system, it is essential that the catalyst should be regenarable and reusable for a number of times. In this context, the regeneration and reuse potential of TiO2 was evaluated. The conversion efficiency of regenerated TiO2 (∼47%) was slightly less compared to the virgin one (55%) for a NOx concentration of 300 ppmv and a RH of 80% (figure not shown). Photocatalytic oxidation is an adsorptiondiffusion process and some of the reactions may be irrevers1038

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 3, 2006

FIGURE 4. Conversion of NOx by ozone in the (a) presence and (b) absence of SO2 (SO2 concentration, 100 ppmv; space time, 10 s; inlet NOx concentration, 200 ppmv; RH, 80%). ible. Permanently blocked sites reduce the surface for adsorption of NO, which in turn decreases the conversion of NOx. The regenerant contained dilute HNO3, which has reuse potential. Thus, the regeneration helps in resource recovery as well. Mass balance for the entire system showed that approximately 72% of the total NOx-N oxidized in the reactor was recovered in the regenerant in the form of nitrous and nitric acids. Oxidation by Ozone. It has been seen that photocatalytic oxidation using TiO2 as catalyst was not able to achieve complete conversion of NO. As an alternative, NO oxidation by ozone was examined. The experiment was conducted at room temperature and atmospheric pressure with a residence time of 10 s and a relative humidity of 80%. NOx concentration of 200 ppm (NO, 175 ppm; NO2, 25 ppm) was employed. Ozone flow was increased gradually, and the oxidation of NO to NO2 was monitored using a flue gas analyzer. It was found that almost all NO was converted (99%) to NO2 when a slight excess (10%) of stoichiometric amount (10.1 mg/ min) of ozone was provided in the gas stream. The results are presented in Figure 4b. The possible reactions which may occur during ozone oxidation of NO are as follows (27)

NO + O3 f NO2 + O2

(i)

NO2 + O3 f NO3 + O2

(ii)

NO2 + NO3 f N2O5

(iii)

NO + NO3 f 2NO2

(iv)

NO + NO3 f 2NO2

(v)

Ozone oxidation of NO at lower temperature (300-600K) occurs largely through the reaction i. Gas phase mass balance of the present study supported this reaction pathway. Effect of SO2 on Oxidation of NO by Ozone. Sulfur dioxide is a common constituent of flue gas along with NOx. To study the effect of sulfur dioxide on oxidation by ozone, the

FIGURE 6. Relative concentrations of nitrate and nitrite in the scrubbed liquid (residence time, 3 s). FIGURE 5. Absorption of NO2 (NO2 concentration, 200 ppmv; residence time, 3 s): (a) effect of flow rates and residence times (scrubbing liquid ) distilled water); (b) effect of different scrubbing different solvents. experiment was carried out with SO2 (100 ppm) in the gas stream, keeping all other reaction conditions the same. Ozone concentration in the gas stream was gradually increased to observe the reaction behavior at different ozone loading rates. The oxidation of NOx was significantly affected by the presence of SO2 (Figure 4a,b). To get the same degree of conversion as in the case when SO2 was not present, much more ozone was consumed. SO2 consumed ozone to get oxidized to sulfate and/or sulfite. This was clear from the sudden decrease in SO2 concentration in the outlet of oxidation chamber. Both SO2 and NO were oxidized completely when an ozone flow rate of 16.3 mg/min was maintained. The loading rates of NO and SO2 to the reactor were 6.3 mg/min and 7.87 mg/min (0.123 mM/min), respectively. Stoichiometric amount of ozone needed for complete conversion of NO to NO2 as per reaction i is 10.1 mg/ min. Ozone left over for SO2 oxidation was 6.2 mg/min (0.13 mM/min). Hence, it can be inferred that 1 mol of ozone was consumed per mole of SO2. The ozone consumption can be reduced by pretreating the flue gas for SO2 using a biotrickling filter or a scrubber as demonstrated by other researchers (15). NO2 capture can be achieved in a second scrubber. Scrubbing. Nitrogen dioxide from the ozone oxidizing reactor was scrubbed at room temperature (27 °C) and atmospheric pressure. Absorption study was performed with different residence times (3, 4, 6, 8 s, based on empty bed) and liquid flow rates (1, 2, 5 and 10 L/h). As expected, absorption increased with increasing EBRT and liquid flow rate as shown in Figure 5a. A flow rate of 5L/h and a contact time of 8 s showed the best results. Absorption of NOx can be explained with the following possible chemical reactions:

2NO2 + H2O f HNO3 + HNO2 3NO2 + H2O f 2HNO3 + NO

∆G° ) -7.75 kcal (vi) ∆G° ) -12.62 kcal (vii)

Absorption studies were also conducted with mineral media, with and without bicarbonate, as scrubbing liquid at an EBRT of 3 s, and the results are shown in Figure 5b. Around 80% of NO2 absorption was noted with a liquid flow rate of 10 L/h for distilled water, whereas it was around 65% and 60% for mineral media with bicarbonate and without bicarbonate, respectively. NO2 is an acidic gas, and it is more absorbable in alkaline media (4). Again, ionic strength of the media plays an important role in the solubility, which on increase gives lower solubility and vice versa. Mineral media with bicarbonate showed slightly higher absorption com-

TABLE 2. Estimated Values of Biokinetic Parameters of Heterotrophic Denitrifiers under Various Operating Conditions leachate from scrubber

synthetic nitrate solution

sample

kinetic paramater

carbon source: sewage

carbon source: ethanol

carbon source: sewage

1 2 3 4

YT kd µmax qmax

0.296 0.06/day 1.9/day 7.8/day

0.476 0.052/day 2.26/day 9.6/day

0.286 0.04/day 1.89/day 7.2/day

pared to mineral media without bicarbonate. This may be attributed to the alkalinity of bicarbonate. Nitrate and nitrite concentrations of leachate were analyzed. It was observed (Figure 6) that, in case of distilled water, the nitrate-nitrite ratio was much higher (∼24) than that of bicarbonate (∼7) media, for all liquid flow rates employed in the present study. From reactions vi and vii, it can be noted that reaction vi (∆G° ) -7.75 kcal) is accompanied with higher standard Gibb’s free energy than reaction vii (∆G° ) -12.62 kcal). Therefore, reaction vii is more favorable, giving a higher ratio of nitrate and nitrite concentration. In the case of bicarbonate media, this effect is partially minimized by ionic strength and alkalinity. Again, solubility of oxygen is higher in distilled water than mineral media, which can convert nitrite to nitrate. As a result of reaction vii, NO is released into the system. This may be the reason for the lower removal efficiency (80-90%) after scrubbing, even though 100% oxidation of NO was taking place by ozone oxidation. Biological Denitrification Study. Denitrification studies were conducted using the enriched heterotrophic denitrifiers. Biokinetic parameters of the consortium were evaluated using ethanol as well as sterilized settled sewage (IIT Madras campus, Chennai, India) as carbon source and electron donors. These results are presented in Table 2. Ethanol seems to be a better electron acceptor compared to sewage, with YT, µmax, and qmax values of 0.476, 2.26/d, and 9.6/d, respectively. Though the sewage system could give only a YT of 0.286 and µmax of 1.89/d, sewage is recommended as the carbon source and electron donor for the system due to its abundance and low cost. The low performance of the system with sewage as electron donor may be attributed to the low BOD/COD (0.4-0.5) ratio of the wastewater. Studies were also conducted to determine the maximum nitrate/nitrite concentration the system can effectively denitrify, for a given concentration of electron donor (COD 450 mg/L). It was found that the nitrate removal efficiency was around 90% up to the nitrate concentration of 800 mg/ L, which corresponded to a COD:N ratio of 2.5:1 (Figure 7), and it dropped to around 80% at a nitrate concentration of VOL. 40, NO. 3, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1039

per ton of NO oxidized. This calculation is based on the following assumptions: (1) the stoichiometric ratio is 10.1 mg/min of O3 for 6.3 mg/min of NO oxidized, (2) 50 W of electricity is required for the production of ozone at the rate of 10 g/h, and (3) the cost of electricity is 9 cents per kwh (kilowatt hour).

Acknowledgments Funding for this project was received from SIDA through a sponsored project on Air Quality Improvement under ARRPET-II.

FIGURE 7. Variation of denitrification efficiency with respect to COD:N ratio (MLSS concentration, 3000 mg/L; hydraulic retention time, 6 h). 1000 mg/L. Maximum specific growth rate (µmax) occurs at a concentration of 1000 mg/L. It has been reported that nitrate removal is greatly dependent on COD:nitrate-nitrogen ratio (29). From the present study, it was observed that the denitrification efficiency was not affected significantly up to a COD:nitrate-nitrogen ratio of 2.5. Denitrification Study Using Scrubbed Leachate. Denitrification kinetic study was carried out with the leachate, for the gas stream NOx concentration of 300 ppm (NO, 260 ppm; NO2, 40 ppm), using sewage as electron donor (figure not shown). Liquid flow rate of 10 L/h and residence time of 3 s were maintained for the scrubbing. The concentration of nitrate in leachate was around 90 mg/L. The denitrification system functioned as a sequential batch reactor with a BSRT of 6 days and an MLSS concentration of 3000 mg/L. The biosystem attained steady state within 15 days of operation, with denitrification efficiency above 90%. The system operated at 6 h reaction time and a settling time of 1 h. Due to the volume constraint of the reactor, the entire leachate was not treated in the system. The leachate volume can be reduced effectively by reducing the scrubbing liquid flow rate. From the kinetic studies, it was found that the denitrification system could perform effectively with efficiency above 90% up to a nitrate concentration of 800 mg/L. The biokinetic constants were almost the same as those of the denitrification system which operated with synthetic scrubbing liquid (Table 2). The lower denitrification rate of the present system (0.14 kg N/kg VSS/day) as compared to the regular denitrification system employing sewage as carbon source (0.06-0.38 kg N/kg VSS/day) may be due to the low quality of sewage. The sewage used in the present study had only a BOD/COD ratio of 0.4-0.5, apart from having traces of heavy metals such as iron and copper (Table 1). Overall, the system presented here was able to remove NOx from flue gases effectively. The end products were environmentally friendly. Since the present system depends on oxidation and scrubbing before the biological process, the high temperature of the flue gases seems to be a boon for the process. It is well established that the rate of photocatalytic and advanced chemical oxidation increases with increase in temperature (24). The treatment option of photocatalytic oxidation followed by ozonation, scrubbing, and denitrification has a definite edge over the one without photocatalytic oxidation, from an environmental point of view. In photocatalytic oxidation process, 72% of the NO oxidized was recovered as nitric acid/nitrous acid. The remaining NO in the gas stream was treated using ozone. Thus ozone consumption was reduced considerably. From an economic point of view, this process requires two additional units, which may considerably increase the capital, operation, and maintenance costs compared to the system with ozone oxidation followed by scrubbing and denitrification. The cost of ozone oxidation is approximately $720 1040

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 3, 2006

Literature Cited (1) Allen, G. C.; El-Tuki, A.; Hallman, K. R.; McLaughlin, D; Stacey, M. Role of NO2 and SO2 in degradation of limestone. Br. Corros. J. 2000, 35, 35-48. (2) Castro, T.; Madronich, S.; Rivale, S.; Muhlia, A.; Mar, B. The influence of aerosols on photochemical smog in Mexico City. Atmos. Environ. 2001, 35 (10), 1765-1772. (3) Fawler, D. Fertilizing the atmosphere with fixed nitrogen, the roles of fossil combustion and agriculture. Institute of terrestrial ecology, Edinburg Research Station. The 22nd Macauly Lecture, May 14, 1998. (4) Wark, K.; Warner, C. F. Air Pollution: Its Origin and Control; Harper and Row Publishers: New York, 1981. (5) Bradford, M.; Grover, R.; Paul, P. Controlling NOx emissions: Part 2. Chem. Eng. Prog. 2002, 98, 42-46. (6) Clean Air Interstate Rule. U.S. Environmental Protection Agency. http:// www.epa.gov/interstateairquality. (7) Pham, E. K.; Chang, S. G. Removal of NO from flue gases by adsorption to an iron (II) thiochelate complex and subsequent reduction to ammonia. Nature 1994, 369, 139-141. (8) Sundaresan, B. B.; Harding, C. I.; May, F. P.; Hendrickson, E. R. Adsorption of nitrogen from waste gas. Environ. Sci. and Technol. 1967, 1, 151-156. (9) Ichiura, H.; Kitaoka, T.; Tanaka, H. Preparation of composite TiO2-Zeolite sheets using a paopermaking technique and their application to environmental improvement. J. Mater. Sci. 2002, 37 (14) 2937-2941. (10) Yan, P.; Fujiwara, M.; Zhou, Y.; Ishida, M.; Investigations on luminescence and NOx renoval by pulse corona discharges. J. Electrostat. 2001, 51-52, 266-271. (11) Izumi, J.; Yasutake, A.; Tomonaga, N.; Tsutaya, H.; Oka, N. Development on high performance gas separation process using gas adsorption. Tech. Rev. 2002, 39, 6-10. (12) Devinny, J. S.; Deshusses, M. A.; Webster, T. S. Biofiltration for Air Pollution Control; CRC-Lewis Publishers: Boca Raton, FL, 1999. (13) Cox, H. H. J.; Deshusses, M. A. Environ. Technol. 2000, 21, 427435. (14) Murali Krishna, N. D. V. S. S.; Philip, L.; Thiobacillus denitrificans immobilized biofilter for NO2 removal. Cleaner Technologies and Environmental Policy. In press. (15) Philip, L.; Deshusses, M. A. Sulfur dioxide treatment from flue gases using a biotrickling filter-bioreactor system. Environ. Sci. Technol. 2003, 37, 1978-1982. (16) Jin, Y.; Veiga, M. C.; Kennes, C. Bioprocesses for the removal of nitrogen oxides from polluted air. J. Chem. Technol. Biotechnol. In press. (17) Flanagan, W. P.; Apel W. A.; Barnes, J. M.; Lee, B. D. Development of gas-phase bioreactors for the removal of nitrogen oxides from synthetic flue gas stream. Fuel 2002, 81, 1953-1961. (18) Davidova, Y. B.; Schroeder, E. D.; Chang, D. Y. P. Proceedings of the 90th Annual Meeting and Exhibition of A&WWA, Toronto, Ontario, Canada, June 8-13, 1997, 97-WP71B.05. (19) Lee, K. H.; Sublette, K. L. Reduction of nitric oxide to elemental nitrogen by Thiobacillus denitrificans. Appl. Biochem. Biotechnol. 1990, 24-25, 441-445. (20) Lee, K. H.; Sublette, K. L. Simultaneous combined microbial removal of sulfur dioxide and nitric oxide from a gas stream. Appl. Biochem. Biotechnol. 1991, 28-29, 623-634. (21) Perry, R. H.; Green, D. W.; Maloney, J. O. Perry’s Chemical Engineers Handbook, 6th ed.; McGraw-Hill International Editions, 1984; pp 3-98. (22) Muruganandham, M.; Swaminathan, M. Photocatalytic decolourisation and degradation of Reactive Orange 4 by TiO2-UV process. Dyes and Pigments 2006, 68, 133-142. (23) Ou, Y.; Lin, J.-D.; Zou, H.-M.; Liao, D.-W. Effects of surface modification of TiO2 with ascorbic acid on photocatalytic

decolorization of an azo dye reactions and mechanisms. J. Mol. Catal. A: Chem. 2005, 241, 59-64. (24) Devahasdin, S.; Fan, C., Jr.; Li, K.; Chen, D. H. TiO2 photocatalytic oxidation of nitric oxide: transient behaviour and reaction kinetics. J. Photochem. Photobiol. A: Chem. 2003, 156, 161170. (25) Dalton, J. S.; Janes, P. A.; Jones, N. G.; Nicholson, J. A. Hallam, K. R.; Allen, G. C. Photocatalytic oxidation of NOx gases using TiO2: a surface spectroscopic approach. Environ. Pollut. 2002, 120, 415-422. (26) Ichiura, H.; Kitaoka, T.; Tanaka, H. Photocatalytic oxidation of NOx using composite sheets containing TiO2 and a metal compound. Chemosphere 2003, 51, 855-860.

(27) Puri, I. K. The removal of NO by low temperature ozone oxidation. Combust. Flame 1995, 102, 512-518. (28) APHA, AWWA, APCF. Standard Methods for the Examination of Water and Wastewater, 1995. (29) Benefield, L. D.; Randall, C. W. Biological Process Design for Wastewater Treatment; Prentice-Hall: Eaglewood Cliffs, NJ, 1982.

Received for review August 1, 2005. Revised manuscript received November 9, 2005. Accepted November 18, 2005. ES0515102

VOL. 40, NO. 3, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1041