The effect of weak acid anions on the selective catalytic wet air oxidation of aqueous ammonia to nitrogen

In this work, the effect of weak acid anions on the ammonia removal has been extensively studied for the process of selective catalytic wet air oxidation (CWAO) of ammonia to nitrogen. It is found that the presence of weak acid anions can effectively enhance the ammonia conversion and selectivity towards nitrogen. The combination between the weak acid anions and H+ to produce weak acid molecules is responsible for such enhancement. Firstly, the H+ consumption of weak acid anions can increase the NH3 concentration and thus the reactivity of ammonia oxidation, due to the shift to NH3 on the equilibrium of NH4 +/NH3. Secondly, the competition combination with H+ between the weak acid anions and NO2 − can increase the concentration of NO2 − and thus boosts the disproportionation reaction between NH4 + and NO2 − to produce nitrogen.

. Supported transition metals and composite oxides have been utilized as the catalysts for decomposing ammonia from wastewater and shown excellent catalytic performance and potential under appropriate reaction conditions [9][10][11][12][13][14][15][16][17][18][19] . Meanwhile, the reaction behavior and mechanism of CWAO of ammonia to N 2 has also been investigated on the purpose of helping to understand the reaction, so as to guide the catalyst development. In order to increase ammonia removal efficiency, strong alkaline were usually added into the feed solution, since ammonia mainly exists as NH 3 in alkaline region and most of studies convinced that NH 3 is much more reactive than NH 4 + on the CWAO of ammonia. Qin et al compared the reactivity of NH 4 + and NH 3 by adjusting the initial pH in the range from 3 to 13 20 . It was found that ammonia oxidation occurred at pH above 10, in which ammonia mainly exists as NH 3 rather than NH 4 + . They thought the N 2 formation was from the catalytic surface reaction, and the NO 3 − was from the further oxidation of NO 2 − . Based on their experiments and investigations, following mechanism was proposed:  2 3 where, *indicates the catalytic active site at vacant state. Lee et al 21,22 also convinced that the reaction of CWAO of ammonia to N 2 proceeded rapidly in strong alkaline region such as pH>10. However, they thought that N 2 was solely produced from the homogeneous reaction between NO 2 − and NH 4 + and the formation of HNO 2 was responsible for the pH decreasing based on the fact that the pH of the solution was decreased from 12 to 2 after reaction. Accordingly, the following reaction pathways partly overlapped the mechanism by Qin et al were proposed: Lousteau et al reported that part of nitrites was reduced to molecular nitrogen via a retro-disproportionation reaction between the ammonium and nitrites ions (NH 4 + + NO 2 − = N 2 + 2H 2 O) while part of them was oxidized into NO 3 − by oxygen. The lower pH (towards acidic condition), the more NO 3 − was produced. They also found that Pd/TiO 2 catalyst can shift the HNO 2 /NO 2 − equilibrium toward the basic form (NO 2 − ) because of high pH (>8), which inhibited one possible evolution route toward the formation of nitrates and in turn "promoted" the retro-disproportionation reaction toward molecular nitrogen formation 12 . There are two key equilibriums in the process for CWAO of ammonia to N 2 : The equilibrium 10 is related to the catalytic oxidation reaction and the equilibrium 11 to the retro-disproportionation reaction between NH 4 + and NO 2 − . It is reasonable to deduce that the addition of the substance that can consume H + in the feed solution will shift the both equilibrium to right. The increasing ammonia concentration will facilitate the oxidation of ammonia and the increasing NO 2 − concentration will favor the formation of N 2 and inhibit the NO 3 − . However, as far as we know, no such work can be found in literature. In this paper, the weak acid anions were introduced into the feed solution to enhance the removal efficiency of ammonia and the selectivity to N 2 . As expected, the addition of weak acid anion can combine with H + and adjust the concentration of NH 3 and NO 2 − during the process of CWAO of ammonia, which effectively improve the catalytic performance.

Experimental Section
Catalyst preparation. Three different kinds of catalysts, Ru-Cu/C, Pt/TiO 2 and Ru/TiO 2 , were synthesized to compare the influence of weak acid anions on the catalytic wet air oxidation of ammonia to nitrogen. H 2 PtCl 6 ·6H 2 O aqueous solution and RuCl 3 ·3H 2 O aqueous solution were used as noble metal precursors, where the concentration were accurately determined by inductively coupled plasma-optical emission spectroscopy (ICP-OES). CuCl 2 .2H 2 O was also used as active metal precursor.
Pt (3 wt%) supported on TiO 2 (Sinopharm Chemical Reagent Co., Ltd. surface area: 10 m 2 /g) were prepared via incipient wetness impregnation method 10 . Ru (3 wt%) supported on TiO 2 (Degussa P-25 Surface area: 50 m 2 /g) and bimetallic catalyst Ru-Cu/C were synthesized by the reduction method described in our previous work 23 . The amount of precursor was adjusted to obtain 1.5% metal loading of Ru and 1.5% metal loading of Cu for the bimetallic catalyst. All the obtained solids were dried at 60 °C in the oven for one night and finally Ru catalysts were calcined at 250 °C and Pt catalyst reduced at 300 °C in flowing hydrogen (4.8 Lh −1 ) for 4 h in the furnace.
Catalyst characterization. The specific surface area of the catalyst was determined by N 2 adsorption (the BET method) using a Micromeritics Tristar 3000 instrument. The chemical composition of the sample was measured by inductively coupled plasma atomic emission spectroscopy (ICP-AES). X-ray powder diffraction (XRD) patterns were collected over a Panalytical X'pert Pro diffractometer operated at 40 kV and 30 mV, using Cu Kα radiation. The images of TEM were obtained on a Tecnai F30 electron microscope (Phillips Analytical) operated with an acceleration voltage of 300 kV.
Apparatus and reaction. The feed solutions for the evaluation of different acid anions were prepared as following: (1) Add 0.55 mL of ammonia solution (mass percentage = 25%) into deionized water to prepare a 100 mL solution containing 1000 ppm ammonia. (2) Take 10 mL solution of (1) and then add the corresponding acid solution drop by drop to adjust the pH to around 8.5. Record the volume of the used acid solution (V). (3) Mix 0.55 mL of ammonia solution (25%) with 10V acid solution, and then deionized water was added into the mixture to 100 mL. The feed solutions for the evaluation of different pH were prepared by dissolving the desired NH 4 Cl or CH 3 COONH 4 into deionized water and then adding 2M NaOH to adjust pH.
The catalytic performance of CWAO of ammonia to N 2 were tested in a 100 mL stainless steel autoclave reactor coated with Teflon liner 2 mm thick inside to avoid corrosion of the reactor wall and a magnetic spin bar was used for stirring. In a typical run, 0.6 g of catalyst and 60 mL reaction solution containing 1000 ppm ammonia was added in the reactor. After adjusting the initial pH to a desired value, the reactor was purged with high purity N 2 five times in order to remove the trace of O 2 . Then, the mixture was heated to the desired temperature under stirring at a rate of 1000 rpm. When the temperature was reached to desired value, 2 MPa air was immediately introduced into the reactor. This moment was considered as zero (starting) time for reaction. Liquid samples were extracted from the reactor at regular intervals of time and filtered by a membrane filter (pore size: 1 μm). After the collected liquid samples were cooled to the ambient temperature, the pH and concentrations of ammonia, NO 3 − and NO 2 − were analyzed. For the reaction behavior between NH 4 + and NO 2 − , 10 mL solution containing the same concentrations of NH 4 + and NO 2 − (500 ppm) was introduced into the reactor. The initial pH of the feed was adjusted to 6.2. Then, the air was introduced and the temperature was increased to 150 °C with a stirring rate of 1000 rpm. 3 hours later, the reactor was separated from the heater and cooled to room temperature. Finally, the liquid sample was treated and analyzed. Product Analysis. The liquid samples were analyzed by Napierian reagent calorimetric method (GBZ5000587) using UV detector at 405 nm to measure the ammonia-nitrogen. The further oxidation products of ammonia (NO 2 − and NO 3 − ) in the liquid phase were measured by HPLC (High Performance Liquid Chromatography) and LC/MS (liquid chromatography/mass spectrometry). The gaseous products were analyzed by a GC equipped with two columns (Porapack T and Molsieve 5A).
As the detected nitrogen containing compounds were only N 2 , NH 3 , NH 4 + , NO 3 − and NO 2 − , the conversion of ammonia (x NH 3 ) and selectivity of nitrogen (S N 2 ) was calculated by the following formulas: where C 0 is the initial concentration of the substrate in the solution, C is its current concentration and the subscripts denote the individual compounds or ions.
The pH values of reaction solution before and after oxidation were measured using a pH meter (Radiometer Analytical PHM240).
Total organic carbon (TOC) content in the liquid samples was measured by TOC-VCSH analyzer which employs high temperature combustion oxidation, coupled with non-dispersive infrared detection technology.

Results and Discussion
Characterization of catalyst. The Pt/TiO 2 , Ru/TiO 2 and Ru-Cu/C catalysts were characterized by ICP, N 2 adsorption, XRD and TEM before catalytic reaction. As shown in Table 1, the loadings of Pt and Ru over Pt/ TiO 2 and Ru/TiO 2 catalysts were 2.86 and 2.89%, respectively. For Ru-Cu/C catalyst, the loadings of Ru and Cu www.nature.com/scientificreports/ were 1.45 and 1.43%. These results indicated that most of active metals were successfully loaded on support as the measured loadings are quite close to the feed. The XRD patters of Pt/TiO 2 , Ru/TiO 2 and Ru-Cu/C catalysts are shown in Figs 1, 2 and 3. For the sake of comparison, the corresponding supports are also shown. The similar XRD patterns of catalysts with supports indicated that the main structure of the supports were retained after the introduction of active metals. Signals for Pt and Ru metals were detected over Pt/TiO 2 and Ru/TiO 2 suggesting that the Pt or Ru species was mainly present in the form of metallic state and was aggregated to some extent, as shown in Fig. 4. No Ru or Cu species was detected over Ru-Cu/C catalyst, which is consistent with our previous study that both Ru and Cu mainly existed in the form of metallic state and was well-dispersed over carbon support 23 . Compared with Ru/TiO 2 , Pt/TiO 2 exhibited more sharp XRD peaks of TiO 2 suggesting that the support for Pt/TiO 2 possessed much better crystal structure with the lowest the specific surface area (as shown in Table 1).
The effect of different acid anions on ammonia conversion. The effects of different acid anions on the ammonia conversion were studied by using ammonium phosphate, ammonium acetate, ammonium sulfate, ammonium chloride and ammonium nitrate as ammonia resource, respectively. The feed solutions were prepared by the procedure described in Experiment section. The experiments were performed at 150 °C and the results are shown in Table 2. It can be seen that the acid anion showed strong effects on the ammonia conversion. Ammonia conversion reached to a value higher than 30% when the feed solution contained weak acid anion. However, for the feed solution containing strong acid anion, only 10% ammonia was converted. To explore the effects of the weak acid anions on the catalytic activity, the reaction behavior of CWAO of ammonium were comparatively investigated by using ammonium acetate and ammonium chloride aqueous solutions in the following study.  The effect of pH on ammonia conversion. The effects of pH on ammonia conversion were studied and the results are shown in Fig. 5. For convenience, the ammonia in ammonium acetate solution was denoted as ammonia-A and that in ammonium chloride solution as ammonia-C. As reported in literature, the initial pH strongly affected the ammonia conversion. The higher pH, the more ammonia was converted. This further convinces that NH 3 is much more active than NH 4 + in the CWAO of ammonia. It can also be seen that the presence of acetate can enhance the reactivity of ammonia, however, such enhancement was vulnerable to pH change. At pH 8.3, ammonia-A showed a conversion of 37.4% while only 3.9% of ammonia-C was converted. With pH  increasing, the gap between conversions of ammonia-A and ammonia-C was narrowed. When pH was higher than 10, no obvious difference between ammonia-A and ammonia-C conversions was observed.
Since NH 3 is more reactive than NH 4 + , one possible reason for the enhancement of ammonia conversion of ammonia-A at low pH was that CH 3 COO − can increase the NH 3 concentration via combining with H + . There are 3 reactions responsible for the H + production during the process for CWAO of ammonia: (1) the dissociation of NH 4 + (NH 4 + ↔NH 3 +H + ); (2) the dissociation of H 2 O (H 2 O↔H + +OH − ); (3) the dissociation of HNO 2 (HNO 2 ↔H + + NO 2 − ). The combination between CH 3 COO − and H + can shift the 3 equilibrium to the right. For reaction (1), the shifting to right can increase the NH 3 concentration. For the reaction (2), the shifting to right increases the OH − concentration, which also increases the NH 3 concentration due to the reaction between the OH − and NH 4 + to produce NH 3 and H 2 O. For the reaction (3), the H + consumption prevents the pH declining and thus increases the NH 3 concentration, which will be proved later in this work.
The effect of temperature on ammonia conversion and N 2 selectivity. Figure 6 shows the effects of reaction temperature on the ammonia conversion and N 2 selectivity of ammonia-A and ammonia-C. When the initial pH was 8.2, the conversion of ammonia-A was always higher than that of ammonia-C and continuously   increased with temperature ascending. When temperature was increased from 150 °C to 200 °C, the conversion was increased from 37.4% to 92.8%. In contrast, ammonia-C kept very low ammonia conversions (<11%) in the whole range of temperature. Temperature had no observable effect on the ammonia conversion for ammonia-C, which is consistent with literatures. M. Bernardi et al reported that the NH 4+ ions were very reactive under moderately acid pH and catalytic wet air oxidation of ammonia can work well at near neutral pH conditions. A high NH 3 conversion was obtained when the reaction was carried out in ammonium acetate aqueous solutions (pH = 6.8) 10 . Another study got more than 90% ammonia decomposition when a phosphate buffered solution (pH = 6.8) was used in order to stabilize the acidic form 24 . However, when ammonium chloride was used as feed solutions, a high ammonia conversion was obtained only until the pH was reached up to 10 20-22 .
Two factors affect the reaction rate of ammonia oxidation reaction: the NH 3 concentration and reaction temperature. The higher NH 3 concentration, the higher reaction rate can be obtained. High temperature favors the ammonia oxidation and enables catalyst to reach a high ammonia conversion. Although the two systems (ammonia-A and ammonia-C) exhibited a similar initial pH, the presence of CH 3 COO − can increase the NH 3 concentration. Therefore, increasing temperature can increase the ammonia conversion of ammonia-A to a great extent, probably due to the available NH 3 . For ammonia-C, NH 4 + is more predominant than NH 3 . Ammonia-C remained a very low ammonia conversion in the whole range of 150 to 200 °C due to the lack of NH 3 , since NH 4 + can be hardly oxidized in CWAO process at temperatures lower than 200 °C.
The effect of temperature on TOC conversion. During the CWAO process, acetate anions can be oxidized to CO 2 . The TOC conversion of ammonia-A as a function of reaction temperature is shown in Fig. 7. It can be seen that TOC conversion kept at a very low level, although it increased with increase of temperature. Even at 200 °C, the TOC conversion was only 25.1%. This result indicated that most of acetate anions were not oxidized and remained in the solution, which was in agreement with the reported data in the literature that acetic acids was refractory to further oxidation. No organic component but CH 3 COO − was detected in the liquid phase after reaction, indicating that the final product of CH 3 COO − oxidation was carbonate ions or carbon dioxide.
Catalytic performance of Pt/TiO 2 and Ru/TiO 2 catalyst. The positive effect of acetate anions on the ammonia conversion was further investigated over another two catalysts, Pt/TiO 2 and Ru/TiO 2 , and the results are shown in Fig. 8. For comparison, Ru-Cu/C catalyst was also included. Similar enhancements for ammonia conversion were observed over the both catalysts, although Pt/TiO 2 , Ru/TiO 2 and Ru-Cu/C catalysts exhibited different specific surface area and active metal particle size. More than 90% ammonia in ammonia-A can be converted over Pt/TiO 2 and Ru/TiO 2 . However, just around 20% and 40% of ammonia was oxidized over Pt/TiO 2 and Ru/TiO 2 respectively, when the feed was switched to ammonia-C aqueous solutions. The pH after reaction is also presented in Fig. 8. It can be seen that, whatever catalyst, the feed solution of ammonia-A always had higher pH after reaction than that of ammonia-C, even although the initial pH were 8.2 for both cases. The possible reason for the higher final pH of ammonia-A is that CH 3 COO − combined with H + to produce CH 3 COOH and prevented the pH decreasing to some extent. Noted that the final pH of ammonia-A was lower than 5.0. Therefore, it is reasonable to think that the final oxidative species of CH 3 COO − should be carbon dioxide, since carbonate ions will be converted into CO 2 under such acidic condition. Figure 9 shows the evolutions of ammonia conversion and pH as a function of time over RuCu/C catalyst. It exhibited much higher ammonia conversions for ammonia-A than the ammonia-C. The ammonia conversion continuously increased with the reaction proceeding and reached a value of 88.7% after 160 minute. For ammonia-C, the ammonia conversion remained very low in the whole process, after 20 min reaction, the ammonia conversion was only 16.1% and the final conversion was 23.9%. In contrast with conversion, the pH decreased as the reaction proceeded in both solutions. In the first 20 minute, the pH decreased more rapidly for ammonia-A and thus was lower than that for ammonia-C. However, after 20 minute, a higher pH was observed for ammonia-A than that for ammonia-C. Generally, the pH decreasing during the CWAO of ammonia is related to the over-oxidation of ammonia to nitrogen oxide rather than nitrogen. At 200 °C, the ammonia is favorably oxidized to nitrous acid. Nitrous acid dissociates into a NO 2 − ion and an H + and thus lowers the pH of the solution. At the beginning of the reaction, more ammonia of ammonia-A was oxidized to nitrite (much higher ammonia conversions) and thus the pH drops down quicker than that for ammonia-C. After 30 min, a higher pH is exhibited for ammonia-A than ammonia-C, although conversion of ammonia in ammonia-A further increased and was much higher than in ammonia-C. These can be explained below. There are two reactions for ammonia consumption: ammonia oxidation and the disproportionation reaction between NH 4 + and NO 2 − . At acidic pH, ammonia is mainly present in the form of NH 4 + and the disproportionation reaction between NH 4 + and NO 2 − was responsible for the ammonia consumption. After 30 min, the pH in ammonia-A was lower than 5.1. Therefore, the conversion of ammonia-A kept increasing while the evolution of pH exhibited a very flat pattern, implying the disproportionation reaction is dominant in this stage when weak acid is presented. For ammonia-C, the flat pattern of pH evolution was contributed to the fact that only very little ammonia was oxidized after 30 min and nearly no disproportionation reaction happened. Very similar results can also be obtained over Pt/TiO 2 and Ru/TiO 2 , as shown in Figs 10 and 11. The evolutions of ammonia conversion and pH were also studied in a phosphate buffered system over RuCu/C catalyst. The results, shown in Fig. 12, indicated that the presence of PO 4 3− can also effectively enhance the ammonia conversion and prevent the solution pH decreasing, similar to CH3COO − ion.

−
. It has been reported that NO 2 − was an intermediate product which can be oxidized to NO 3 − during the process of CWAO of ammonia at 200 °C. As shown in Fig. 10, the decreasing pH indicated that part of ammonia-A was oxidized to nitrite, however, the N 2 selectivity is higher than 98% (shown in Fig. 6) and only very little nitrate was detected. Therefore, the effect of acetate anion on the behavior of the disproportionation reaction between NH 4 + and NO 2 − and nitrite oxidation reaction was investigated and the results are shown in Table 3. For comparison, the effect of pH was also included.
As can be seen, pH strongly affected the two reactions. When pH was at 10.5, the reaction between NH 4 + and NO 2 − scarcely occurred, due to the very low percentage of NH 4 + in alkaline environment and no detectable nitrate was formed. This implied that NO 2 − can not be oxidized to NO 3 − , since the nitrite is predominantly present in the form of NO 2 − at pH higher than 10. At pH 6.2, more than 63% of ammonia and most of the nitrite  were converted. However, more than 30% of nitrite was oxidized to NO 3 − . With further decreasing pH to 3.6, the ammonia consumption reaction was restricted to some extent and more than 50% of ammonia was left due to the scarcity of NO 2 − , since HNO 2 is becoming more dominant to NO 2 − at a lower pH. In contrast with ammonia consumption reaction, the reaction of nitrite oxidation to NO 3 − was further enhanced. The concentration of NO 3 − reached a high value of 245 ppm, implying HNO 2 was oxidized to NO 3 − . Acetate anion exhibibt a strong influence on the nitrite oxidation reaction, although solution 2 and 4 (in Table 3) exhibited very comparable ammonia conversion. For solution 2, most of nitrite was converted, and some was reacted with NH 4 + to form N 2 and part of them was oxidized to nitrate. For solution 4, quite a lot of nitrite was left and only 70 ppm nitirate was dectected. This result indicated that part of nitrite in solution 4 neither   reacted with ammonia (169 ppm was left) nor was oxidized to nitrate. The fact of little production of nitrate indicate that the nitrite in the solution 4 was maily present in the form of NO 2 − , not HNO 2 , although the pH was relatively acidic (6.2). One knows that the equilibrium constant of CH 3 COOH is much lower than that of HNO 2 . Therefore, it is reasonable to deduce that the equilibrium of CH 3 COO − /CH 3 COOH consumed H + and made nitrite present in the form of NO 2 − . Another question is why quite an amount of ammonia was left and did not react with NO 2 − to produce N 2 . This can be explained that the equilibrium of CH 3 COO − /CH 3 COOH made some ammonia present in the form of NH 3 , which can not react with NO 2 − . This was consistent with the fact the much higher reactivity of ammonia oxidation was attribute to the higher concentration of ammonia at acidic pH.

Conclusion
The catalytic behavior for CWAO of ammonia to N 2 was comparatively investigated in various ammonium salt solutions over Ru-Cu/C, Pt/TiO 2 and Ru/TiO 2 catalysts. It was found that the weak acid anion could increase the NH 3 concentration in the solution and thus enhance the reactivity of catalytic oxidation of ammonia. The weak acid anion can also increase the concentration of NO 2 − and facilitate the retro-disproportionation reaction between NH 4 + and NO 2 − toward N 2 . The both factors contribute the improving ammonia removal efficiency and/or the selectivity to N 2 .