Environ. Sci. Technol. 2002, 36, 5462-5468
Selective Photocatalytic Oxidation of NH3 to N2 on Platinized TiO2 in Water
JAESANG LEE, HYUNWOONG PARK, AND WONYONG CHOI* School of
Environmental Science and Engineering, Pohang University of Science and Technology, Pohang 790-784, Korea
Selective photocatalytic oxidation of NH3 to N2 is proposed as a new treatment method for controlling the levels of ammonia in water. The photocatalytic oxidation of ammonia on naked and metallized TiO2 in water saturated with air, nitrogen, or N2O gas was investigated. While the slow photocatalytic oxidation of NH3 to NO2-/NO3- is the only pathway for decomposition of NH3 on naked TiO2 and Au/ TiO2, a new pathway, that of selective oxidation of ammonia to dinitrogen, opens up on Pt/TiO2. The formation of dinitrogen from the oxidation of 15NH3 was confirmed by mass spectrometric detection of 15N2. The photocatalytic conversion of NH3 to N2 greatly increases when the Pt/ TiO2 suspension is saturated with N2O gas, whereas N2O itself shows little reactivity with naked TiO2 and Au/TiO2. Over 80% of the total nitrogen available in ammonia (0.1 mM) is converted into N2 within 40 min illumination of the N2Osaturated Pt/TiO2 suspension. The ability of N2O to accept the conduction band electrons of Pt/TiO2 was verified by photoelectrochemical measurements. N2O reductively decomposes to generate OH radicals on Pt/ TiO2; the rate of ammonia degradation in the N2O-saturated Pt/TiO2 suspension significantly decreases in the presence of excess tert-butyl alcohol, an OH radical scavenger. The presence of Pt deposits on the TiO2 particles changes the photocatalytic pathway of ammonia conversion by both enhancing OH radical production from N2O and stabilizing intermediate NHx (x)0,1,2) species to facilitate their recombination into N2.
Ammonia is one of the major nitrogen-containing pollutants in wastewater; it is a source of nutrients that may accelerate the eutrophication of and algal growth in natural waters and is a common product of the chemical/biological transformation of organic nitrogenous pollutants (1). High concentrations of ammonia in wastewater effluents deplete dissolved oxygen, reduce chlorine disinfection efficiency, and exhibit acute toxicity to aquatic life. The World Health Organization also recommends that the total amount of ammonia in drinking water should not exceed 1.5 mg/L since it causes a disagreeable taste and smell at trace levels. To cope with the deterioration of water quality produced by the discharge of ammonia into aquatic systems, several chemical and physical methods have been developed and applied in the field such as air stripping, ion exchange, and breakpoint
* Corresponding author phone: +82-54-279-2283; fax: +82-54279-8299; e-mail: firstname.lastname@example.org.
chlorination (1, 2). Among these methods, the biological nitrification/denitrification process is generally regarded as the most efficient method for remediating high concentrations of ammonia in wastewater (1). The increasing interest in advanced oxidation processes (AOPs) has encouraged researchers to investigate their applicability to the removal of ammonia from water (3-10). In the 1970s, the ozonation process was proposed as a method for oxidation of ammonia in wastewater (5, 6). A few studies also reported the photocatalytic oxidation of ammonia using titanium dioxide (3, 4, 8-10). These oxidation processes, which involve hydroxyl radicals as a primary oxidant, transform ammonia into nitrite/nitrate quantitatively. However, since nitrite/nitrate is more toxic than ammonia, a better remediation process would be to convert ammonia into dinitrogen. In fact, there are electrochemical oxidation processes that have been reported as converting ammonia into dinitrogen on platinum electrodes (11, 12). Semiconductor photocatalysis has been made use of in a variety of chemical conversion systems (13). While much research has been directed toward achieving nitrogen fixation (N2fNH3) using photocatalysis as a part of solar energy conversion technology (14-16), the reverse photocatalytic conversion has not been reported. Since most photocatalytic remedial action is enabled by the strong oxidizing power of OH radicals (10, 17, 18), it is very difficult to attain selectivity in TiO2 photocatalytic reactions. In the case of the oxidation of ammonia, the problem is how to convert ammonia [N(-III)] selectively to dinitrogen [N(0)], not to completely oxidized nitrate [N(+V)]. Among the many possible methods of modifying photocatalyst reactivities, metallization has been frequently tried. Deposition of noble metals such as Pt, Pd, Au, and Ag on semiconductor surfaces enhances their photocatalytic efficiency by retarding electron-hole recombination (19, 20). Semiconductors coated with these metal deposits have been used successfully in various systems for remediation of polluted water and air and have exhibited high efficiency in photocatalytic reduction processes (19-23). The presence of metal deposits not only enhances photocatalytic activity but also in general changes reaction pathways. In this work, we demonstrate that NH3 can be selectively converted to N2 on platinized TiO2 under UV illumination. Pt deposits on TiO2 stabilize the intermediate atomic nitrogen species and thus facilitate their recombination into dinitrogen. The conversion of NH3 to N2 is particularly effective when the system is saturated with N2O. In contrast to biological ammonia conversion which is a two-step process (nitrification/denitrification), the present photocatalytic transformation proceeds in a single step.
Chemicals and Materials. NH4Cl (Aldrich), 15NH4Cl (Aldrich), NaNO2 (Aldrich), NaNO3 (Aldrich), N2H4 (Aldrich), NH2OH (Aldrich), HAuCl4 (Sigma), H2PtCl6 (Aldrich), and tert-butyl alcohol (t-BuOH) (Shinyo) were used as received. For the analysis of hydroxylamine levels, FerroZine reagent, NH4(CH3CO2), and Fe(ClO4)3 were purchased from Aldrich. The gases (air, Ar, N2, and N2O) used were of > 99% purity. Air, Ar, and N2 were obtained from BOC-Gases and N2O was obtained from Dongbang Inc. The water used was ultrapure (18 M??cm) and prepared with a Barnstead purification system. Titanium dioxide (Degussa P25), a mixture of 80% anatase and 20% rutile, was used as the photocatalyst. Other chemicals used were of the highest purity available.
10.1021/es025930s CCC: $22.00 ? 2002 American Chemical Society Published on Web 11/08/2002
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 24, 2002
Preparation of Metallized TiO2. Metallization of TiO2 was carried out using a photodeposition method. Platinization was carried out by irradiating (with a 200 W medium-pressure mercury lamp) an aqueous suspension of TiO2 (0.5 g/L) for 30 min in the presence of 1 M methanol (electron donor) and 0.1 mM chloroplatinic acid (H2PtCl6). After irradiation, the filtered Pt/TiO2 sample was collected and washed with distilled water. The concentration of unused chloroplatinic acid remaining in the filtrate solution after photodeposition was determined by inductively coupled plasma-atomic emission spectroscopy (ICP-AES, Thermo Jarrell Ash Cooperation IRIS/AP) in order to quantify the amount of deposited Pt. A typical Pt loading on the TiO2 particles was estimated to be ca. 0.2 wt %. Au was photodeposited from chloroauric acid onto the TiO2 particles by following a similar procedure. Transmission electron micrographic images of Pt/TiO2 showed that the resulting Pt particles were in the size range 1-2 nm and were well dispersed on the TiO2 particles (2030 nm). Photolysis and Analysis. All naked or metallized TiO2 suspensions were prepared in water at a concentration of 0.5 g/L and were dispersed by simultaneous sonication and shaking for 30 s in an ultrasonic cleaning bath. The substrate NH3 stock solution (10 mM) was added to the suspensions to make the desired concentration (typically 0.1 mM), and then the pH of the suspensions was adjusted to pH 10.0 with 1 M NaOH standard solution. For the experiments in the presence of various dissolved gases such as air, Ar, N2, and N2O, the reactor was continuously purged with the corresponding gas. Assuming saturation at room temperature, the concentration of dissolved O2 (under air-saturation), N2, and N2O in the suspensions was 0.26, 0.66, and 24 mM, respectively. Loss of ammonia due to volatilization and adsorption onto TiO2 surfaces was minimal within the time scale of the present experiments. After gas bubbling for 30 min, photoirradiation was commenced using a 300 W Xe-arc lamp (Oriel). The light was passed through a 10 cm IR water filter and a UV cutoff filter (λ > 300 nm). The filtered light was focused onto a 200 mL Pyrex reactor with a quartz window. The photocatalytic reactor was filled with minimized headspace, sealed with a rubber septum, and stirred magnetically. Sample aliquots of 1.5 mL were withdrawn from the illuminated reactor with a 2 mL syringe, filtered through a 0.45 ?m PTFE filter (Millipore), and injected into a 4 mL glass vial. More than a duplicate set of photolysis experiments were carried out for a given condition. Quantitative analyses of the ionic intermediates and products were performed using an ion chromatograph (IC, Dionex DX-120). The IC system was equipped with a Dionex IonPac AS-14 for detection of anions, a Dionex IonPac CS12A for detection of cations, and a conductivity detector. The possible formation of NH2OH as an intermediate was checked colorimetrically using a FerroZine method in which hydroxylamine reduces Fe(III) species to Fe(II) species whose complexes with FerroZine can be detected by visible light absorption (24, 25). Aliquots (each of 0.5 mL) of Fe(III) (0.5 mM), ammonium acetate buffer, and FerroZine (10 mM) were mixed with 3.5 mL of sample solution. After 1 h of color development, the concentration of NH2OH in the samples was determined by monitoring the absorbance at 562 nm. The detection limit for NH2OH with the FerroZine method is less than 10 ?M, according to the literature (25). GC/MS Analysis for Detecting 15N2. Qualitative identification of the N-containing gas products evolved from the ammonia oxidation was carried out by using gas chromatography/mass spectrometry (GC/MS). To distinguish atmospheric N2 from the dinitrogen produced by the photocatalytic oxidation of ammonia, 15NH4Cl (98% atom 15N, Aldrich) was used as the source of ammonia. A much higher ammonia concentration (1 mM) was employed for this
purpose. A Pyrex reactor with a quartz window and a sampling port was purged with ultrapure Ar gas for 30 min prior to the photolysis, and then the sampling port was sealed. The gas trapped in the reactor headspace during the photolysis was sampled by a 500 ?L gastight syringe and analyzed by injection into a GC/MS. The GC/MS system consisted of a gas chromatograph (HP 6890) equipped with a HP-5DB column and a mass selective detector (HP 5973 MSD). The electron impact energy for ionization was 70 eV. Photoelectrochemical Measurements. The effects of the purging gases on the photoelectrochemical characteristics of the TiO2 and Pt/TiO2 electrodes were compared. TiO2 electrodes were prepared as described previously (26). One milliliter of TiO2 suspension (5 wt %) was spread over indium tin oxide (ITO) glass (1 × 1 cm2; Samsung) and dried. This TiO2/ITO electrode was then heated at 450 °C for 30 min. For Pt deposition onto a TiO2 electrode, the electrode was immersed in an aqueous solution of 0.4 g/L H2PtCl6 and 0.1 M methanol and illuminated with a 200 W Hg lamp for 2 h. The photoelectrochemical reactor was cylindrical and had a working electrode (TiO2 or Pt/TiO2), a reference saturated calomel electrode (SCE), a counter Pt-gauze electrode, and a gas injection port. The electrolyte used was distilled water or 0.1 mM NH4Cl at pH 10 and the purging gas was O2, N2, or N2O. The UV light source was three 10 W black-light lamps (Sankyo Denki) and the UV intensity was 0.13 mW?cm-2 at 30 W. Open-circuit potentials and short-circuit currents were measured by a potentiostat (EG&G, Model 263A) that was connected to a computer.
Results and Discussion
The Photocatalytic Degradation of NH3 under Air and N2 Saturation. The removal of ammonia and the formation of intermediates and products during the course of photocatalytic degradation on naked and metallized TiO2 in airsaturated conditions are compared in Figure 1. Ammonium ions (NH4+) have little reactivity with OH radicals (27, 28) and do not degrade at all in the UV/TiO2 system (3, 4, 8-10), whereas neutral ammonia degrades on TiO2 photocatalytically (4, 8-10). Therefore, all the photolysis experiments in the present work were carried out at pH 10-11 (above ammonia’s pKa)9.3). The photocatalytic degradation of ammonia on naked TiO2 (Figure 1a) is slow and results in its stoichiometric conversion to NO2- and NO3-, which agrees with previous results (8-10). The total N-mass balance could be satisfactorily accounted for throughout the photolysis, which indicates that there were no other products. The slight decrease in total-N shown in Figure 1a,b could be due to volatile loss or adsorption of ammonia. Au/TiO2 showed a similar reactivity to naked TiO2, except that NO2- production was higher than that of NO3- (Figure 1b). On the other hand, the photocatalytic degradation of ammonia on Pt/TiO2 (Figure 1c) was much faster and was accompanied by a significant reduction in the total N-mass, which implies the presence of other products. Hydrazine (N2H4) might be produced through the dimerization of the amino radicals (NH2?) that form from the reaction of NH3 with OH radicals. Although N2H4 has not been reported as an intermediate of ammonia degradation in photocatalytic systems (8-10), its formation has been suggested by a pulse radiolytic study of the oxidation of NiII(NH3)n complexes by NH2? radicals (27). In this study, however, IC analysis detected no sign of hydrazine formation. The presence or otherwise of hydroxylamine (NH2OH) was also checked according to the method described in the Experimental Section but was produced only at concentrations beneath the detection limit (<10 ?M), if at all. Figure 2 shows the results of the photocatalytic degradation of ammonia under N2 saturation. In the absence of dissolved oxygen, NH3 degradation on naked TiO2 was
VOL. 36, NO. 24, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
FIGURE 1. Photocatalytic conversion of NH3 (C0 ) 0.1 mM) with (a) naked TiO2, (b) Au/TiO2, and (c) Pt/TiO2 under air-equilibrated conditions. The irradiation was started at time “zero” after 30 min equilibration. The ordinate scale, C/C0, refers to concentrations normalized with respect to the initial ammonia concentration. The reactor was open to the air during irradiation. The loss of NH3 through volatilization was insignificant. Other experimental conditions were pH ) 10; [TiO2] ) 0.5 g/L. negligible, whereas its degradation was slightly enhanced on Au/TiO2 and much enhanced on Pt/TiO2. It is interesting to note that the kinetics of ammonia oxidation on metallized TiO2 (for both Au/TiO2 and Pt/TiO2) showed little dependence on whether dioxygen was present (compare Figures 1 and 2). Ammonia oxidation on Pt/TiO2 in the anoxic suspension also showed a significant mass deficit in the total N balance.
FIGURE 2. Photocatalytic conversion of NH3 (C0 ) 0.1 mM) with (a) naked TiO2, (b) Au/TiO2, and (c) Pt/TiO2 under N2-saturated conditions. The suspension was continuously purged with N2 during irradiation. The loss of NH3 through volatilization was insignificant. Other experimental conditions were the same as those for Figure 1. This implies that gaseous products such as N2, NO, and N2O are generated as a result of NH3 degradation. The Photocatalytic Degradation of NH3 under N2OSaturation. Figure 3 shows the influence of N2O on the photocatalytic oxidation of ammonia on naked TiO2, Au/ TiO2, and Pt/TiO2. Nitrous oxide, which scavenges the aqueous electrons produced by the γ-radiolysis of water (27, 29), has often been used as an alternative to dioxygen as a conduction band (CB) electron acceptor in semiconductor
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 24, 2002
TiO2 and Au/TiO2 (Figure 3a,b). However, N2O on Pt/TiO2 produced a dramatically enhanced effect (Figure 3c): the NH3 in the Pt/TiO2 suspension disappeared completely within 40 min, and less than 20% of the total NH3 depleted was converted into NO2-/NO3-, which reassured the presence of gaseous products generated on Pt/TiO2. The different reactivities of N2O with the TiO2 and Pt/TiO2 surfaces have been previously recognized: decomposition of N2O into N2 and O2 does not occur on naked TiO2 but does on Pt/TiO2 (36). The high efficiency of N2O as an electron acceptor should be responsible for the complete removal of NH3 on Pt/TiO2. We checked the possibility that N2O decomposition might produce NO2-/NO3- by carrying out an N2O photolysis in a Pt/TiO2 suspension in the absence of NH3. No NO2- and NO3- were detected throughout the N2O photolysis. Any further UV irradiation beyond the point of complete removal of NH3 increased the concentration of NO3- (Figure 3c) only marginally, which indicates that no other intermediate that could be oxidized to NO3- was produced in the N2O-Pt/ TiO2 system. Most of the ammonia removed by photocatalysis on Pt/TiO2 seemed to have been converted into gaseous products. Analysis of Gas Products by GC/MS. In the mass spectra obtained from the GC/MS analysis we detected a peak at m/e ) 30 (15N2), produced by the photocatalytic oxidation of 15NH3 on Pt/TiO2. A similar experiment using naked TiO2 did not produce any m/e ) 30 signal. Mass peaks corresponding to other gaseous nitrogen compounds such as 15NO and 15N2O also could not be detected. Although a direct quantification of the amount of evolved 15N2 could not be achieved in this experiment, a relative abundance comparison of the 15N2 generated from the photocatalytic oxidation of 15NH on Pt/TiO and the 15N produced by a chemical 3 2 2 oxidation of 15NH3 was carried out. The chemical oxidant used was hypobromite (OBr-), which has been reported as transforming ammonia into dinitrogen through the following reaction (37):
215NH3 + 3OBr- f 3Br- + 3H2O + 15N2
FIGURE 3. Photocatalytic conversion of NH3 (C0 ) 0.1 mM) with (a) naked TiO2, (b) Au/TiO2, and (c) Pt/TiO2 under N2O-saturated conditions. Other experimental conditions were the same as those for Figure 1. photocatalytic reactions (30-36). N2O has been shown to scavenge CB electrons trapped on ZnO (30-34) and TiO2 (34, 35). However, N2O has been reported to have little activity on TiO2 surfaces as an electron acceptor with respect to enhancing the self-decomposition of N2O (36) and to the photocatalytic degradation of organic compounds (33, 34), although it has shown activity in photocatalytic reactions on ZnO surfaces (30-34). In accordance with these previous observations, we found that N2O activity differed little from N2 activity in inducing the photooxidation of NH3 on naked
Hypobromite was prepared by adding bromine to a solution of lithium hydroxide. Under conditions similar to those of the photocatalytic experiment, the oxidation of 15NH3 was initiated by injecting excess hypobromite into an ammonia solution after 30 min of Ar gas purging. The IC analysis showed that the complete chemical oxidation of the ammonia was achieved without formation of ionic products such as NO2- and NO3-. GC/MS analysis of the headspace gas after the completion of chemical oxidation also detected a mass peak of m/e ) 30, whose intensity was a little higher than that obtained from the photocatalytic oxidation of 15NH on Pt/TiO . This confirms that most of the NH on 3 2 3 Pt/TiO2 was photocatalytically transformed into N2 in this study. Since the nitrite, nitrate, and dinitrogen are the observed products of the photocatalytic conversion of ammonia, the N-mass deficits shown in Figures 1c, 2c, and 3c should be ascribed to dinitrogen. Based on this fact, the conversion efficiency of NH3 to N2 on Pt/TiO2 after 2 h irradiation is estimated to be 65-70% in air- or nitrogensaturated suspensions and over 80% under N2O saturation. The Effects of Pt on the Dinitrogen Formation Mechanism. While group VIII metals on semiconductors are known to trap CB electrons and subsequently to enhance interfacial electron transfers through a Schottki barrier, their catalytic effects on photoinduced reactions should also be considered. Platinum has been used as a catalyst in various heterogeneous reactions such as hydrogen-deuterium exchange, hydrogenation of ethylene, and oxidation of carbon monoxide (20, 38). The transition metals such as Pt, Pd, and Rh have been reported to have a stronger affinity for ammonia than the
VOL. 36, NO. 24, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
coinage metals (Cu, Ag, and Au) (12). Computed adsorption energies of atomic nitrogen on noble metals have shown that the Pt surface (-394 kJ mol-1) has an optimal moderate atomic nitrogen affinity for the production of dinitrogen in comparison with the affinities of Ru (-525 kJ mol-1) and Rh (-448 kJ mol-1), on which N atoms are too tightly bound to recombine; Au (-162 kJ mol-1) and Ag (-156 kJ mol-1) bond with N atoms too weakly to produce active intermediates (12). Due to this unique property of Pt as a catalyst, ammonia can be transformed into dinitrogen on platinum electrodes by electrochemical oxidation (11, 12): NH3 adsorbed on a Pt electrode is oxidized to NHx radical species (x ) 1, 2) that are subsequently transformed into N2 through their diffusion and recombination. In accordance with this scheme, N, NH, and NH2 radical species formed by thermally activated oxidation of NH3 have been shown using surface spectroscopic techniques to be stabilized on the Pt (111) and Pt (100) surface (39, 40). On the other hand, a similar electrochemical oxidation study (12) reported that NH3 conversion to N2 was inactive on Cu, Ag, and Au electrodes, which agrees with our observation that Au/TiO2 is inactive. Judging from our results and these previous findings, dinitrogen formation seems to be enabled by the stabilization of the intermediate NHx species on Pt deposited on TiO2. Based upon an electrochemical mechanism proposed by Gerischer and Mauerer (11), the following reaction steps are suggested here as a plausible mechanism for the selective photocatalytic conversion of NH3 to N2 on Pt/TiO2. However, it should be noted that the photocatalytic mechanism could be different from the electrochemical mechanism and thus needs further experimental verification.
TiO2 + hv f ecb- + hvb+ (charge-pair generation) ecb + Pt f Pt(ecb ) (electron trapping)
hvb+ + H2O (>OHad) f ?OH (OH radical generation) (4) NH3,aq f NH3,ad (NH3 adsorption) NH3,ad + ?OH f NH2,ad(Pt) + H2O
NH2,ad(Pt) + ?OH (or hvb ) f NHad(Pt) + H2O (or H ) (7) NHad(Pt) + ?OH (or hvb+) f Nad(Pt) + H2O (or H+) (8) NHx(Pt) + NHy(Pt) f N2Hx+y(Pt) (x,y ) 0,1,2) N2Hx+y(Pt) + (x+y)hvb+ f (9)
FIGURE 4. Dark adsorption of NH3 on naked TiO2, Au/TiO2, and Pt/TiO2. (a) The adsorptive removal of NH3 (C0 ) 0.1 mM) on naked TiO2, Au/TiO2, and Pt/TiO2 as a function of the equilibration time and (b) the adsorption isotherms of NH3 on naked TiO2, Au/TiO2, and Pt/TiO2 ([NH3]ad measured after 24 h equilibration). The control experiments were done in the absence of photocatalysts in order to estimate the volatile loss. The experimental conditions were pH ) 10; [TiO2] ) 0.5 g/L; air-equilibrated. mM, it becomes significant at higher concentrations (Figure 4b). The isotherms of naked TiO2 and Au/TiO2 are almost identical with the control data, which indicates that the adsorption of NH3 onto naked TiO2 and Au/TiO2 is negligible. On the other hand, Pt/TiO2 exhibits a much higher adsorption affinity for NH3. The adsorption of ammonia onto Pt/TiO2 proceeded gradually over 6 h and further equilibration produced little change (Figure 4a). The slow adsorption kinetics of ammonia on Pt/TiO2 implies that the photocatalytic conversion process could be limited by diffusion effects. The markedly decelerating rates of ammonia removal shown in Figures 1c and 2c reflect the fact that the initial surface concentration of ammonia on Pt/TiO2 was rapidly depleted but only slowly replenished from the bulk aqueous phase. Although the present photolysis experiments were performed after 30 min equilibration, a much longer preirradiation equilibration should yield higher conversion rates of ammonia on Pt/TiO2. The adsorption isotherm of ammonia on Pt/TiO2 (Figure 4b) shows no sign of saturation up to [NH3]0 ) 1 mM, which implies that all the experiments in this work were performed at surface concentrations of NH3 that are far below monolayer coverage.
N2,ad + (x+y)H+ (N2 formation) (10) (11) (12)
N2,ad f N2,aq (N2 desorption) NH3,ad + ?OH f HONH2,ad(Pt) + Had(Pt) HONH2,ad(Pt) + ?OH ff NO2 ff
NO3- (NO2-/NO3- formation) (13)
Gerischer and Mauerer (11) proposed that the active intermediates for dinitrogen formation were NHx (x)1,2) species and that the formation of Nad deactivated the electrode surface. The formation of active intermediates NHx on the naked TiO2 surface is not favored so reaction 13 should be dominant, as we observed. The favorable adsorption of nitrogen species such as NH3 and NHx (x)0,1,2) onto the Pt surface is a prerequisite for the selective photocatalytic oxidation of NH3 to N2. Figure 4 shows that NH3 adsorption characteristics are notably different between naked TiO2, Au/TiO2 and Pt/TiO2 surfaces. Although the volatile loss of NH3 is negligible at [NH3]0 ) 0.1
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 24, 2002
TABLE 1. Effects of Purging Gases on the Open-Circuit Voltage (EOC) and Short-Circuit Current (ISC) in a Photoelectrochemical Cell with TiO2 or Pt/TiO2 Electrode
electrode/ electrolyte TiO2/H2Oc TiO2/NH4Cld Pt-TiO2/H2Oc Pt-TiO2/NH4Cld O2 -51 -187 -39 -21 ?EOC (mV)a N2 N2 O -252 -322 -120 -163 -230 -310 -32 -56 ?ISC (?A/cm2)b O2 N2 N2O 22.4 21.3 3.6 4.0 39.4 34.3 8.0 10.1 27.3 33.6 3.3 5.4
a ?E b ?I OC ) illuminated EOC - dark EOC. SC ) illuminated ISC - dark ISC. c Distilled water at pH 10. d 0.1 mM NH4Cl at pH 10.
The Role of N2O on Pt/TiO2. In terms of energetics, N2O should not be a better electron acceptor than O2 since the electron affinity (EA) of the former is much lower than that of the latter [EA(N2O) ) 0.22 eV; EA(O2) ) 0.45 eV] (41). Although Anpo et al. (35) verified that N2O could trap CB electrons of TiO2 by detecting N2O?- in an electron spin resonance study, its subsequent decomposition into N2 and O2 did not take place on naked TiO2 (36). In this study, N2O did not enhance the photocatalytic oxidation of NH3 on naked TiO2 and Au/TiO2. On the other hand, the fact that N2O can degrade on Pt/TiO2 (36) is in accord with our results. However, the highly enhanced photocatalytic conversion of NH3 on Pt/TiO2 in the presence of N2O should not be ascribed to the electron scavenging action of N2O because the ammonia removal efficiency with Pt/TiO2 is lower in the presence of the more powerful electron acceptor, dioxygen. To confirm the ability of N2O to accept CB electrons from Pt/TiO2, we compare the steady-state open-circuit voltages (EOC) and short-circuit currents (ISC) in a photoelectrochemical cell with a TiO2 or Pt/TiO2 electrode in O2-, N2-, or N2Osaturated solutions in Table 1. If the dissolved gases can accept CB electrons from the electrode surface, the voltage (or current) difference between the illuminated and dark condition, ?EOC (or ?ISC), should decrease. Table 1 shows that both ?EOC and ?ISC for the TiO2 electrode under O2saturation were much smaller than for N2- and N2Osaturation, which implies that O2 readily accepts CB electrons from the TiO2 electrode whereas N2O and N2 do not. On the other hand, the Pt/TiO2 electrode under N2O-saturation exhibited comparable ?EOC and ?ISC values to those for O2 saturation, indicating that N2O accepts CB electrons from Pt/TiO2 to a similar degree to O2. The most plausible role of N2O in this system is the production of OH radicals on Pt/TiO2. On the Pt surface, N2O accepts CB electrons (reaction 14) and subsequently dissociates to yield OH radicals (reactions 15-17)
FIGURE 5. Photocatalytic removal of NH3 on Pt/TiO2 in the presence or absence of t-BuOH. The experimental conditions were [NH3]0 ) 0.1 mM; [TiO2] ) 0.5 g/L; [t-BuOH]0 ) 10 mM; N2O-saturated. inhibited the removal of NH3 from the N2O-saturated Pt/TiO2 suspension. This confirms that N2O serves as a precursor of OH radicals on Pt/TiO2. The heterogeneous electron scavenging ability of N2O seems to sensitively depend on the kind of semiconductor surface that is used. N2O scavenged trapped electrons from a ZnO surface in a similar way to O2 (30-34) and assisted the photodecomposition of organic compounds on ZnO (34), whereas N2O is much less efficient in scavenging CB electrons from TiO2. Serpone et al. (33) speculated that trapped electrons on TiO2 might be much more strongly bound than those on ZnO. The presence of Pt on TiO2 could release trapped electrons from tightly bound TiIII sites into the platinum phase where electrons are much more mobile. As a result, trapped electrons on Pt could be easily transferred to N2O. The Pt deposits on TiO2 enable the selective oxidation of NH3 to N2 in two ways: (1) by trapping CB electrons and reductively dissociating N2O into OH radicals that initiate NH3 oxidation and (2) by stabilizing intermediate NHx species on Pt and facilitating their recombination to dinitrogen. Using a Pt/TiO2 suspension may not in general be economical as a practical water treatment technology. Catalyst immobilization or recovery processes should be considered for practical applications. It should be also noted that this selective photocatalytic conversion of ammonia to N2 is applicable only to NH3, not to NH4+. The present method could be viable when dealing with low concentration levels of ammonia (less than a few ppm) in water.
N2O + ecb-(Pt) f N2O?-(Pt) N2O?-(Pt) f N2 + O?N2O?-(Pt) + H2O f N2 + OH- + ?OH O?-(Pt) + H2O f OH- + ?OH
(14) (15) (16) (17)
This work was supported by POSCO and partly by the Brain Korea 21 program.
The resulting hydroxyl radicals react with NH3 to generate N2 through reactions 5-11. To verify the involvement of OH radicals in the oxidation of NH3 in the N2O-saturated suspension, the effect of adding excess t-BuOH, an OH radical scavenger (10), was investigated (Figure 5). Given the reported rate constants and concentrations employed [k (t-BuOH + OH?) ) 5 × 108 M-1s-1 (42), k (NH3 + OH?) ) 1.7 × 107 M-1s-1 (27); [t-BuOH]0 ) 10-2 M, [NH3]0 ) 10-4 M], it is expected that all OH radicals should be scavenged by t-BuOH. As shown in Figure 5, the presence of excess t-BuOH significantly
(1) Tchobanoglous, G.; Burton, F. L. Wastewater Engineering: Treatment, Disposal and Reuse, 3rd ed.; McGraw-Hill: New York, 1991. (2) Water Quality and Treatment: A Handbook of Community Water Supplies; Pontius, F. W., Ed.; McGraw-Hill: New York, 1990. (3) Schmelling, D. C.; Gray, K. A. Water Res. 1995, 29, 2651. (4) Klare, M.; Scheen, J.; Vogelsang, K.; Jacobs, H.; Broekaert, J. A. C. Chemosphere 2000, 41, 353 (5) Singer, P. C.; Zilli, W. B. Water Res. 1975, 9, 127. (6) Hoigne, J.; Bader, H. Environ. Sci. Technol. 1978, 12, 80.
VOL. 36, NO. 24, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
(7) Kuo, C.-H.; Yuan, F.; Hill, D.-O. Ind. Eng. Chem. Res. 1997, 36, 4108. (8) Bravo, A.; Garcia, J.; Domenech, X.; Peral, J. J. Chem. Research(S) 1993, 376. (9) Bosen, E.; Schroeter, S.; Jacobs, H.; Broekaert, J.-A. C. Chemosphere 1997, 35, 1431. (10) Kim, S.; Choi, W. Environ. Sci. Technol. 2002, 36, 2019. (11) Gerischer, H.; Mauerer, A. J. Electroanal. Chem. 1970, 25, 421. (12) de Vooys, A. C. A.; Koper, M. T. M.; van Santen, R. A.; van Veen, J. A. R. J. Electroanal. Chem. 2001, 506, 127. (13) Hoffmann, M. R.; Martin, S. T.; Choi, W.; Behnemann, D. W. Chem. Rev. 1995, 95, 69. (14) Schrauzer, G. N.; Guth, T. D. J. Am. Chem. Soc. 1977, 99, 7189. (15) Rusina, O.; Eremenko, A.; Frank, G.; Strunk, H.-P.; Kisch, H. Angew. Chem., Int. Ed. Engl. 2001, 40, 3993. (16) Tennakone, K.; Wickramanayake, S.; Fernando, C. A. N.; Lleperuma, O. A.; Punchihewa, S. J. Chem. Soc., Chem. Commun. 1987, 1078. (17) Cho, S.; Choi, W. J. Photochem. Photobiol. A: Chem. 2001, 143, 221. (18) Choi, W.; Hong, S. J.; Chang Y. S.; Cho, Y. Environ. Sci. Technol. 2000, 34, 4810. (19) Mill, A.; Le Hunte, S. J. Photochem. Photobiol. A: Chem. 1997, 108, 1. (20) Photocatalysis-Fundamentals and Applications; Serpone, N., Pelizzetti, E., Eds.; Wiley-Interscience: New York, 1989. (21) Einaga, H.; Futamura, S.; Ibusuki, T. Environ. Sci. Technol. 2001, 35, 1880. (22) Ranit, K. T.; Viswanathan, B. J. Photochem. Photobiol. A: Chem. 1997, 108, 73. (23) Tada, H.; Teranishi, K.; Ito, S. Langmuir 1999, 15, 7084. (24) Stookey, L. L. Anal. Chem. 1970, 42, 779. (25) Bourke, G. C. M.; Stedman, G.; Wade, A. P. Anal. Chim. Acta 1983, 153, 277.
(26) (27) (28) (29) (30) (31) (32) (33) (34) (35) (36) (37) (38) (39) (40) (41) (42)
Park, H.; Kim, K. Y.; Choi, W. J. Phys. Chem. B 2002, 106, 4775. Lati, J.; Meyerstein, D. Inorg. Chem. 1972, 11, 2393. Zellner, R.; Smith I.-W. M. Chem. Phys. Lett. 1974, 26, 72. Staehelin, J.; Buhler, R. E.; Hoigne, J. J. Phys. Chem. 1984, 88, 5999. Tanaka, K.; Blyholder, G. J. Phys. Chem. 1971, 75, 1037. Wong, N.-B.; Taarit, Y. B.; Lunsford, J. H. J. Chem. Phys. 1974, 60, 2148. Kamat, P. V.; Patrick, B. J. Phys. Chem. 1992, 96, 6829. Serpone, N.; Maruthamuthu, P.; Pichat, P.; Pelizzetti, E.; Hidaka, H. J. Photochem. Photobiol. A: Chem. 1995, 85, 247. Sukharev, V.; Kershaw, R. J. Photochem. Photobiol. A: Chem. 1996, 98, 165. Anpo, M.; Aikawa, N.; Kubokawa, Y. J. Chem. Soc., Chem. Commun. 1984, 644. Kudo, A.; Sakata, T. Chem. Lett. 1992, 2381. Feast, N. A.; Dennis, P. F. Chem. Geol. 1996, 129, 167. Richardson, J. T. Principles of Catalyst Development; Plenum Press: New York, 1989. Bradley, J. M.; Hopkinson, A.; King, D. A. J. Phys. Chem. 1995, 99, 17032. Mieher, W. D.; Ho, W. Surf. Sci. 1995, 322, 151. CRC Handbook of Chemistry and Physics; David, R. L., Ed.; CRC Press: New York, 1996; pp 10-189. Staehelin, J.; Hoigne, J. Environ. Sci. Technol. 1985, 19, 1206.
Received for review July 2, 2002. Revised manuscript received September 22, 2002. Accepted October 9, 2002. ES025930S
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 24, 2002