Environ. Sci. Technol. 2003, 37, 5745-5749
Mechanism and Kinetics of the Catalytic Oxidation of Aqueous Ammonia to Molecular Nitrogen DEUK KI LEE* Division of Civil and Environmental Engineering, Gwangju University, Gwangju 503-703, Korea
Aqueous phase catalytic oxidation of ammonia has been studied over Ru/TiO2 catalyst in a batch reactor by changing the solution pH, concentration of catalyst in the solution, temperature, and reaction time. The oxidation reaction of ammonia over Ru/TiO2 catalyst has been found to take place exclusively for the aqueous NH3 with a preferred mode in strong alkaline pH region. An oxidation reaction pathway has been proposed as follows: Oxidation of ammonia is initiated by the reaction of aqueous ammonia with catalytically activated oxygen. After undergoing further successive oxidation reactions with activated oxygen, ammonia is finally oxidized to a molecule of nitrous acid. Nitrous acid dissociates into a nitrite ion and a proton. The solution pH is decreased with the protons from the dissociation of HNO2 so that the solution concentration of NH4+ is increased. Molecular nitrogen as a final product is produced from the homogeneous aqueous phase reaction between nitrous ion and ammonium ion. Further reaction of nitrous ion with the activated oxygen leads to the formation of nitrate ion. The reaction pathway proposed has been validated with the changes of solution pH along with the ammonia conversions, and the formation of N2 from the solution containing NO2- and NH4+ ions in equimolar amounts of nitrogen has been confirmed in a separate experiment. The kinetics of aqueous ammonia oxidation reaction has been well represented as a first-order reaction with respect to the concentration of aqueous ammonia, and an apparent rate constant has been obtained as a function of catalyst concentration in solution, oxygen pressure, and reaction temperature.
Introduction Due to the worldwide problem of eutrophication in lakes and rivers, the abatement of ammonia in wastewaters discharged has become a prime issue of environmental control. Because ammonia is largely used as an indispensable chemical in various industries from manufacturing fertilizers to electronics, the risk of environmental contamination by ammonia escaping as a gas or as an aqueous species is increasing. Although biological treatment is known as the most economic and efficient method for removal of contaminants from wastewaters, it is not generally applicable for industrial wastewaters of toxic nature (1). Such chemical or physical treatment technology as breakpoint chlorination or air stripping is available for removal of aqueous ammonia (2). * Corresponding author phone: +82-62-670-2394; fax: +82-62670-2192; e-mail:
[email protected]. 10.1021/es034332q CCC: $25.00 Published on Web 10/22/2003
2003 American Chemical Society
Recently, catalytic wet air oxidation is showing potential as an alternative method for the removal of ammonia from industrial wastewater, in which ammonia can be removed by conversion mainly to molecular nitrogen (3-8). The catalytic wet oxidation of aqueous ammonia stems from the wastewater treatment technology for the wet oxidation of high COD wastewaters with water-dissolved oxygen under high pressure at elevated temperatures. Although catalysts are required to basically alleviate the severity of the reaction condition, as far as the removal of aqueous ammonia is concerned, the conversion of ammonia and the selectivity of N2 formation against NO3- as final products are reported to be dependent on the kind of catalysts used (3, 7, 8). Typical catalysts attempted for ammonia removal in the reaction are precious metals (Pt, Pd, Rh, Ru) or base metals (Fe, Ni, Mn, Co, Cu, W, V) supported on various supports. It has been reported that precious metal catalysts are more active and selective for ammonia removal and for N2 formation, respectively, than base metal catalysts (7, 8). Base metal catalysts are reported to have low activity for ammonia oxidation and favor the formation of NO3- over N2 (7). However, there is an inconsistency in the reports for some precious metal catalysts in the respects of their catalytic activity and selectivity in the reaction. While one reports that the alumina-supported catalyst of Pt is less active and selective than the Ru or Pd (7), there is a report that the Pt catalyst of titania-support is superior to the others (8). There has been no answer about what a kind of role the catalyst plays in the selective production of N2. In general, it has been suggested that the water needs to be alkalinized above pH 9 for the effective oxidation reaction of aqueous ammonia (4, 7), but a report that the reaction is performed under near neutral conditions without alkalization of the aqueous solution is also available (8). Water-contained ammonia exists as predominantly aqueous NH3 in basic solutions above pH 9.3, but as NH4+ in the solutions below pH 9.3. The result that the reaction proceeds under near neutral water in which ammonia as NH4+ predominates leads to a conclusion that NH4+ ions are required for the species to be oxidized. This is a contradiction to the studies (4, 7) that the water needs to be alkalinized for the reaction. There are still many unknowns about the catalytic ammonia oxidation reaction in aqueous phase, particularly with respect to the kind of aqueous nitrogen-species involved in the oxidation reaction, the reaction mechanism leading to the final products N2 or NO3-, and the role of catalyst in the reaction. In this study, reaction experiments for the aqueous phase oxidation of ammonia has been conducted over Ru/TiO2 catalyst in a batch reactor by changing the solution pH, concentration of catalyst in the solution, temperature, and reaction time. By quantitative measurements of the concentrations of all the reactants and products, and the pH of solutions in each experiment, the reaction pathway and the reaction kinetics of the catalytic aqueous ammonia oxidation have been proposed.
Experimental Section Materials. A ruthenium (3 wt %) supported on TiO2 (Degussa P-25, surface area: 50 m2/g) catalyst was used in the reaction. Impregnation of Ru onto the powered support was conducted using an aqueous solution of RuCl3 (Aldrich) in a rotary evaporator. After drying the Ru-impregnated sample at 110 °C for 12 h, the sample was oxidized in a muffle furnace at 500 °C for 6 h. The calcined samples in powder were reduced under a flow of pure H2 at 350 °C for 4 h and stored in a vial for use in reaction. All the solutions were prepared using VOL. 37, NO. 24, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5745
purified water. NH4Cl (Aldrich), NaNO2 (Aldrich), and NaNO3 (Aldrich) were used to make aqueous solutions of ammonia, nitrite, and nitrate, respectively. He (99.999%, Hankook Gas) and O2 (99.99%, Hankook Gas) were used for experiments. Reaction Experiments. All reactions were carried out in a stainless steel reactor inside which a Teflon liner 1.5 mm thick was placed to avoid corrosion of the inside reactor wall. The total volume of the reactor was 220 mL. In a typical batch experiment, 1 g of catalyst and a magnetic spin bar for stirring were placed in the reactor, and 100 mL of feed solution in ammonia concentration of 1000 ppm as NH3 was poured into the reactor. When the initial pH of the feed solution needed to be changed, a few drops of 10 M NaOH were added to the feed sample during the pH measurement, and the feed was immediately poured into the reactor. The ambient air initially present in the reactor was purged with O2 by repeating the injection of pure O2 at 5 bar followed by venting out. After repeating the oxygen purge four times the N2 present in the reactor was analyzed to be negligible. Initial pressure of reactor at room temperature was adjusted with the O2 during a fifth injection. Thereafter, the reactor was heated to a desired reaction temperature, at which stirring with a rate of 1400 rpm using a magnetic bar was started to initiate the reaction. After scheduled reaction time passed, the reactor was separated from heater and quenched in a bath of ice-water. As the reactor temperature fell to room temperature, the gaseous products were vented to a collection bottle graduated for the measurement of gas volume. About 30 mL of the reactor supernatant liquid was quickly sampled to a vial and stored in acidified state by adding a few drops of 1 M HCl for the analysis of ammonia concentration. The other liquid sample from the reactor was collected in a beaker to measure the solution pH and subsequently filtered using a membrane filter (pore size: 1 µm) to separate the liquid from the catalyst. This liquid sample was stored in a sealed bottle for the analyses of nitrite and nitrate concentrations. Analysis. The pH and the concentration of ammonia were measured using a pH electrode (Fisher Accumet) and an ammonia gas sensing electrode (Corning, 476130), respectively, both connected to a pH/ISE meter (Fisher AR25). For the measurement of NH3 concentration, 10 M NaOH was used as an ionic strength adjustor. The measurement of the concentrations of nitrite and nitrate ions was conducted using a HPLC (Vintage 2000) equipped with a conductivity detector. IonPac AS4A-SC column (Dionex) in combination with IonPac AG4A-SC guard column was used for the separation of the anions, and 4 mM carbonate/1.5 mM bicarbonate was used for eluent solution. The gas sample was taken from the gas collection bottle using a gastight syringe and analyzed by a gas chromatograph (Younglin M600D) equipped with a thermal conductivity detector. A column of Molecular Sieve 5A in series with Porapak Q was used to separate possible gas components, O2, N2, NO, and N2O. The amount of sample gas injected into the column was maintained constant 0.5 mL using a 6-port valve. The amounts of gas components produced throughout reaction were determined quantitatively using the GC-analyzed concentrations and the volume of gas collected.
Results and Discussion Effects of Reaction Variables. Results of the oxidation reaction at 200 °C for 5 h for the feed solution with different initial pH under the O2 pressure of 5 bar at room temperature are listed in Table 1. The conversion of ammonia is limited to 20% for the feed of pH 5.5 without alkalization, but higher conversions are attained for the feeds of higher initial pH. N2 is produced as the major gas product, but NO and N2O are not detected. As a liquid product, nitrate ion is detected, and its production is higher for the feeds of higher initial pH. 5746
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 24, 2003
TABLE 1. Results of the Ammonia Oxidation Reactiona over Ru/TiO2 Catalyst for the Feed Solutions with Different Initial pH pH before reaction
convb (%)
N2
5.5 8.1 9.1 10.5 12.3
20.0 41.4 57.6 93.5 99.7
17.3 37.9 52.4 89.5 94.4
yield (%)c NO2- NO30.0 0.0 0.0 0.0 0.0
2.6 3.3 4.3 5.1 7.6
%N balanced
pH after reaction
99.9 99.8 99.1 101.1 102.3
2.3 2.2 2.2 2.5 2.9
a Reaction conditions: NH , 1000 ppm; catalyst, 1 g; initial P , 5 bar 3 O2 (at 25 °C); 200 °C; 5 h. b 100 × (mol NH3 disappeared)/(mol NH3 present initially). c Yield of N2 ) 100 × 2 × (mol N2 formed)/(mol NH3 present initially). Yield of NO2- or NO3- ) 100 × (mol NO2- or NO3- formed)/ (mol NH3 present initially). d 100 × (2 × mol N2 + mol NO2- + mol NO3-)/(mol NH3 present initially).
FIGURE 1. Changes of NH3 conversion and N2 yield (A) and nitrite and nitrate concentrations (B) with reaction time (O2 pressure charged at 25 °C, 5 bar; temperature, 200 °C). The equilibrium relationship between the aqueous ammonia and the ammonium ion is as follows:
NH4+ S NH3(aq) + H+, pKA ) 9.25 at 25 °C, ∆H° ) 12.42 kcal/mol (1) The equilibrium relationship and the results of Table 1 indicate that the oxidation reaction of ammonia over the Ru/TiO2 catalyst takes place exclusively for the aqueous NH3 with a preferred mode in strong alkaline pH region. For the solution of pH 5 in which only 0.02% of total NH3-N is estimated to be present as aqueous ammonia at room temperature, the observed 20% of ammonia conversion may be attributed to the increased presence of aqueous ammonia at 200 °C by virtue of the equilibrium shift toward the dissociation of ammonium ion. Because reaction 1 is an endothermic process, the equilibrium is expected to shift toward the aqueous ammonia formation as the temperature rises, resulting in a higher fraction of aqueous ammonia in the solution at 200 °C than at room temperature. Thereafter, the pH’s of all the feed solutions have been adjusted to 12.212.4 for the oxidation reaction. Figure 1 shows the reaction time-dependent behaviors of ammonia conversion and N2 yield (upper) and the concentrations of nitrites and nitrates (lower) in the reaction at 200 °C under the O2 pressure of 5 bar at room temperature. Ammonia converts rapidly in the first hour to N2, NO2-, and
FIGURE 2. Changes of NH3 conversion and N2 yield (A) and nitrite and nitrate concentrations (B) with the gram weight of catalyst used (temperature, 200 °C; O2 pressure charged at 25 °C, 5 bar; time, 1 h).
FIGURE 4. Changes of NH3 conversion and N2 yield (A) and nitrite and nitrate concentrations (B) with the O2 pressure at the reaction temperature of 200 °C (O2 pressures charged at 25 °C: 3, 5, 7, 10 bar, respectively; time, 1.5 h). in the production of nitrates in a higher concentration. Figure 4 shows the results of the reaction at 200 °C for 1.5 h depending on the pressure of O2 at a reaction temperature of 200 °C. The values of pressure (4.8, 7.9, 11.1, and 15.9 bar) on the abscissa of Figure 4 have been obtained using the ideal gas law from the pressures (3, 5, 7, and 10 bar) of O2 charged initially to the reactor of the gas-phase volume of 110 mL at room temperature. During the reaction at 200 °C, pressures of the reactor were measured as 15.5, 19, 20.5, and 25 bar. Figure 4 indicates that higher pressures of O2 favor the formation of nitrates instead of N2. Reaction Pathway. The above reaction results are significant in helping us establish the reaction pathway. First, the nitrogen-species involved in the oxidation reaction is aqueous ammonia, and consequently the feed solution needs to be alkalinized for the effective reaction. Second, the reaction proceeds over the catalyst. Third, nitrite ions are the reaction intermediates. Last, as shown in Table 1, the pH’s of the product solutions are lower than those of the feeds. Based on these findings, a reaction pathway in the catalytic oxidation of ammonia can be proposed as follows
FIGURE 3. Changes of NH3 conversion and N2 yield (A) and nitrite and nitrate concentrations (B) with reaction temperature (O2 pressure charged at 25 °C, 5 bar; time, 1 h). NO3-. The concentration of nitrate ion is stabilized at about 280 ppm, but that of nitrite reaches a maximum at 1 h, and as the reaction time is extended to about 2 h, it approaches zero. This is the reason the nitrite is not detected in the experiments in Table 1 where the reaction time is 5 h. The time-dependent concentration profile of nitrite ion suggests it is responsible for the reaction forming N2 and NO3- as reaction intermediates. Figure 2 shows the results of the reaction for 1 h depending on the weight of the catalyst charged to the reactor. Ammonia conversion and the yields of N2, NO2-, and NO3- are observed largely dependent on the amounts of catalyst loaded. When no catalyst is used, no conversion of ammonia is attained in practice, indicating that the oxidation of aqueous ammonia takes place only in the presence of catalyst at 200 °C. Figure 3 shows the results of the reaction depending on the reaction temperature. At the reaction for 1 h, a higher conversion of ammonia and a higher yield of N2 are attained for a higher reaction temperature. Particularly, a higher reaction temperature results
O2(aq) + 2* S 2O*
(2)
NH3(aq) + O* f NH* + H2O
(3a)
NH* + O* S NHO* + *
(3b)
NHO* + O* f HNO2 + 2*
(3c)
HNO2 S H+ + NO2-, pKA ) 4.5 at 25 °C
(4)
NO2- + NH4+ f N2 + 2H2O
(5)
NO2- + O* f NO3- + *
(6)
where * indicates the catalytic active site. At first, waterdissolved oxygen adsorbs on the catalyst and dissociates into the activated oxygen O*. Oxidation of ammonia is initiated by the reaction of aqueous ammonia with the activated oxygen, as shown in reaction 3. After undergoing further successive oxidation reactions with the activated oxygen, ammonia is oxidized to a molecule of nitrous acid. Nitrous VOL. 37, NO. 24, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5747
FIGURE 5. Comparison of the solution pH after reaction between the measured and the calculated for the reaction data presented in Figure 1. acid dissociates into a nitrite ion and a proton. As more protons are produced, the solution pH decreases, which makes the equilibrium of reaction 1 shift toward the formation of NH4+. Molecular nitrogen as a final product is produced from the homogeneous aqueous phase reaction between nitrous ion and ammonium ion, as shown in reaction 5. Further reaction of nitrous ion with activated oxygen leads to the formation of nitrate ion. In the reaction scheme above, all the elementary reactions except reaction 5 are the same as those proposed by Qin and Aika (7). Qin and Aika (7) proposed that the N2 formation results from the catalytic reaction between NH* and NHO* on the catalyst. The validity of the reaction pathway proposed can be tested by comparing the pH of the solutions measured after reaction with the pH expected by the reaction scheme and by ascertaining if reaction 5 takes place. First, the pH change in the solution throughout the reaction can be estimated, on the basis of molar yields of N2, NO2-, and NO3- from the initial aqueous ammonia, as follows: By eq 3, 1 mol of NH3 produces 1 mol of HNO2. Nitrite ions dissociated from HNO2 can be converted to N2 or NO3- by reaction 5 or 6, respectively. The mol number of nitrites converted to N2 and NO3- can be calculated from the molar yields of N2 and NO3- produced. Protons from the HNO2 dissociation are responsible for the decrease in the solution pH, but parts of them, being used to transform aqueous NH3 to NH4+ via reaction 1, have no effect on the pH. So, the net mol number of protons in charge of a change in the product solution pH is equivalent to the total number of protons produced minus those consumed forming NH4+. The total number of protons produced in the solution can be obtained by summing the molar amounts of NO2- and NO3- remaining in the solution and those of NH4+ and NO2used to form gaseous N2 by the reaction 5. Because a half molar amount of N2 formed results from NH4+, the net mol number of protons affecting the pH of the resultant solution can be calculated as follows:
mol H+net ) molNO2- + (molN2)/2 + molNO3- (7) Therefore, the pH of the solution of volume 100 mL after the reaction is given by
pH or pOH ) -log
(
)
|molH+net - molOH-init| 0.1
(8)
where mol OH-init indicates the mol number of hydroxide ions present initially in the feed solutions with pH adjusted to 12.2-12.4. Figure 5 shows the pH of the solutions after the reaction at the same condition as in Figure 1, in comparison with those calculated by eqs 7 and 8. The result is similar to 5748
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 24, 2003
FIGURE 6. Reactor pressure increment with temperature (A) for the solution containing NO2- (14.7 mmol N) and NH4+ (14.7 mmol N) in comparison with pure water and the cumulative amount of N2 produced (B) from the reaction between NO2- and NH4+ with temperature (pressure of He charged at 25 °C, 5 bar).
FIGURE 7. Data fit by the first-order reaction kinetics for the effects of temperature (A) and the O2 pressure (B) at the reaction temperature, T in legends (catalyst charged, 1 g). a neutralization curve appearing in the titration of strong base with strong acid. The point of equivalence is present at about 80% conversion of ammonia, at which both lines coincides, validating the proposed reaction scheme. Next is to ascertain the aqueous phase homogeneous reaction between NO2- and NH4+ leading to N2. A reaction was conducted for a 100 mL solution in which NO2- and NH4+ ions are present in equimolar amounts of atomic nitrogen (14.7 mmol N). If the nitrogen-species in the solution is completely converted to N2 by eq 5, there will be N2 production amounting to 14.7 mmol in the gas phase. Without
-rNH3 ) kappCNH3
FIGURE 8. Calculated vs experimental ammonia concentrations. adding the catalyst, the reactor containing the solution was purged and charged with He at 5 bar (instead of O2). The reactor was heated under stirring to 200 °C at a rate 6 °C/ min, during which the pressure of a reactor with an increasing temperature was read out. At 200 °C, the reactor was quenched, and the reactor-vented gas was measured by volume and analyzed with GC. The same experiment was conducted for 100 mL of pure water. Figure 6(A) shows the reactor pressures measured with increasing temperature. The difference of pressures between the solution containing both NO2- and NH4+ and the pure water corresponds to the amount of N2 produced at each temperature. The total amount of N2 produced from the homogeneous reaction was analyzed to be 14.1 mmol, representing a N2 yield, 95.9%. Figure 6(B) shows the cumulative amount of N2 produced with temperature. The reaction of N2 formation is accelerated above 135 °C and nearly completed at 175 °C. This indicates that reaction 5 is very rapid in kinetics. As a result, the yield of N2 produced appears as usual much higher than that of NO3-. Reaction Rate. As far as the reaction for the feed solutions with pH 12.2-12.4 is concerned, the factors affecting the reaction rate are the pressure of O2, temperature, and the catalyst concentration in the solution. To exclude a possible effect of oxygen mass transfer limitation on the reaction rate, the stirring rate of the magnetic bar in the reaction was kept at 1400 rpm as there appeared only minor differences in the conversion of ammonia with the rate of stirring above 1100 rpm. The pressure of O2 in the reactor decreases during the reaction as the O2 is consumed. In most experiments conducted under the initial O2 pressures of 5 bar and above, the decreases throughout the reaction were less than about 14% as compared with those initially charged. Under the assumption that the O2 pressure is not changed throughout the reaction, the rate of disappearance of ammonia can be written as follows
(9)
where rNH3 is the reaction rate of aqueous ammonia oxidation (ppm/h), kapp is the apparent rate constant (1/h), and CNH3 is the concentration of aqueous ammonia in solution. Figure 7 shows the conversion data fit to the first-order kinetics above. Apparent rate constants represented by the slopes of the straight lines in Figure 7 are dependent not only on temperature (Figure 7(A)) but also on the O2 pressure (Figure 7(B)). The data shown in Figure 7 are obtained from the experiments charged with 1 g of catalyst. As shown in Figure 2, the kinetic rate is also dependent on the change in the concentration of catalyst in the reactants. Accordingly, the three affecting factors should be incorporated into the apparent rate constant, which is represented by
ln kapp ) ln A + R lnWc + β ln PO2 -
EA 1 R T
()
(10)
where A is the frequency factor (1/h), Wc is the concentration of catalyst in the solution (g/L), PO2 is the pressure of O2 at reaction temperature, EA is the activation energy (kcal/mol), R is the gas constant, and T is the temperature (K). Throughout the regression for eq 10 using all the data presented in Figures 1-4, the values of A, R, β, and EA are evaluated as 3.661 × 109, 1.1, 1.1, and 24.8, respectively. Figure 8 shows the concentration of ammonia calculated by eqs 9 and 10 in comparison with the experimental concentrations, resulting in a fair consistency.
Acknowledgments This work was supported by the Korea Research Foundation Grant (KRF-2001-002-E00172).
Literature Cited (1) Lee, Y. W.; Ong, S. K.; Sato, C. Water Sci. Technol. 1997, 36, 69-74. (2) Reynolds, T. D.; Richards, P. A. Unit Operations and Processes in Environmental Engineering, 2nd ed.; PWS Pub. Co.: 1996. (3) Barbier, J., Jr.; Oliviero, L.; Renard, B.; Duprez, D. Catal. Today 2002, 75, 29-34. (4) Ukropec, R.; Kuster, B. F. M.; Schouten, J. C.; van Santen, R. A. Appl. Catal. B: Environ. 1999, 23, 45-57. (5) Deiber, G.; Foussard, J. N.; Debellefontaine, H. Environ. Pollut. 1997, 96(3), 311-319. (6) Huang, T. L.; Macinnes, J. M.; Cliffe, K. R. Water Res. 2001, 35(9), 2113-2120. (7) Qin, J.; Aika, K. Appl. Catal. B: Env. 1998, 16, 261-268. (8) Taguchi, J.; Okuhara, T. Appl. Catal. A: Gen. 2000, 194-195, 89-97.
Received for review April 10, 2003. Revised manuscript received August 30, 2003. Accepted September 12, 2003. ES034332Q
VOL. 37, NO. 24, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5749
