Simultaneous removal of NO_X and SO₂ with H₂O₂ over Fe based catalysts at low temperature

Abstract

Simultaneous flue gas desulfurization and denitrification are achieved with ˙OH radicals generated from H₂O₂ catalyzed by Fe based transition metal (Fe, Fe–Al, Fe–Ti, Fe–Si) oxides in the duct, followed by purification with ammonia-based glass made washing tower. The main recovered products are the ammonia nitrate and sulfate, which are important fertilizers. In this oxidation-removal process, 100% SO₂ and 74% NO_X removal efficiencies are achieved at the simulated flue gas temperature of 120–180 °C under the applied conditions with a space velocity of 36 000 h⁻¹. The experimental results show that the enhancement of Lewis acid benefits the generation of O₂ without ˙OH radical production. The increase of Fe³⁺ bonded with hydroxyl group and points of zero charge value are beneficial to produce ˙OH radicals, thus improving the NO_X removal efficiency.

---

Introduction

Nitrogen oxides (NO_X) and sulfur dioxide (SO₂) cause serious air pollution, reduce air quality, and threaten human life. The combined wet flue-gas desulfurization (WFGD) and selective catalytic reduction (SCR) is the mostly used desulfurization and denitrification process.¹ However, this combined process requires high investment and running cost. Therefore, researches on simultaneous removal of NO_X and SO₂ technology have attracted a growing number of attentions.^2,3 Among these processes, the plasma process is an effective, low-temperature and uncontaminated one to purify the flue gas, but the high running cost limits its industrial application.⁴ Other simultaneous removal processes also suffer from problems of low-efficiency or high running or investment cost. Therefore, it is desirable to develop a low-cost, high-efficiency simultaneous removal process for the SO₂ and NO_X.

Some transition metal oxides have been verified to catalytically decompose H₂O₂ to ˙OH radicals, such as Fe₂O₃, Co₂O₃ or ZrO₂ etc.^5–7 It has been known that ˙OH radicals are highly reactive.⁷ NO_X and SO₂ should be oxidized immediately to nitric acid (HNO₃) or sulfuric acid (H₂SO₄) with the injection of ˙OH radicals. Both H₂O₂ and transition metal oxides are of low cost and available raw materials (the economic calculation is shown in the ESI, S11†). Thus, a new simultaneous removal process for NO_X and SO₂ is developed as show in Fig. 1. ˙OH radicals from the catalyzed decomposition of H₂O₂ by Fe based transition metal oxides (results of EPR characterization and experiments have shown that ˙OH radicals do generate under the catalysis of Fe based catalysts, S1 and S2†) are injected into the duct under the purge of nitrogen (N₂). NO_X and SO₂ are partially oxidized into HNO₃ and H₂SO₄ mists, respectively. Finally, the oxidized simulated flue gas is further purified with ammonia-based glass made washing tower. In this paper, the H₂O₂ flow ranges from 0.3 to 0.5 mL min⁻¹. The flow of purged gas is 40 mL min⁻¹, and the total flow is 240 mL min⁻¹. NO_X content is 550 ppm, and SO₂ content is 17 000 ppm. The catalyst in the fix-bed is 0.6 g. The space velocity is 36 000 h⁻¹.

Fig. 1 The combined process on removal of SO₂ and NO_X.

(The experimental and characteristic part is shown in ESI, S10†).

Effect of catalyst supports on removal efficiency

Surface properties have been proved to significantly affect the decomposition of H₂O₂ into ˙OH radicals.^8–11 And catalyst supports have an obvious influence on surface properties. So here some experiments were conducted with different catalyst supports and the results are shown in Fig. 2.

Fig. 2 NO_X and SO₂ removal efficiencies over Fe, Fe–Al, Fe–Ti and Fe–Si. Reaction conditions: 0.6 g catalyst, [NO_X] = 550 ppm, [SO₂] = 17 000 ppm, total flow rate = 240 mL min⁻¹, flow rate of free radicals carrier gas = 40 mL min⁻¹, H₂O₂ flow rate = 0.40 mL min⁻¹, GHSV = 36 000 h⁻¹.

It could be seen that SO₂ achieved nearly 100% removal. However, only 50% NO_X could be removed for Fe–Si. As the catalyst support is changed to titania, iron oxides or alumina, the NO_X removal efficiency increases from 65% to 74%. Moreover, when the temperature increases from 120 °C to 180 °C, the SO₂ removal efficiency remains more than 98%, but the NO_X removal efficiency increases slightly from 64% to 74%. Constant volume of the absorbent in glass made ammonia-based washing tower were injected into the ion chromatograph (with Fe as catalyst). The type of ammonia compounds was determined from its peak retention time in the chromatogram and the compound concentration was calculated from the peak area using an external standard method. In the experimental sample, the peak areas of NO₂⁻, NO₃⁻, SO₃²⁻ and SO₄²⁻ were 0, 2.76, 140 and 1.40, respectively. These results show that ˙OH radicals prefer to oxidize NO_X to HNO₃ mists due to the low activation energy of oxidation reaction.³ (The characterization of ion chromatograph is shown in ESI, S3†) The material balances for NO_X and SO₂ were calculated as shown in Table 1. Results show that no other byproducts will be produced in this oxidation-removal process with Fe based catalysts.

Table 1 Material balance summary for NO_X and SO₂

Category	Run 1	Run 2
Time t, min	60	120
Gas flow Q, mL min^−1	800	800
Solution volume VL, mL	600	600
Cin (NO)	451	451
Cout (NO)	157	160
NO removal efficiency, %	60.8	60.1
Cin (SO2), ppm	17 000	17 000
Cout (SO2), ppm	15	18
SO2 removal efficiency, %	100	100
C(NO2^−) actual value, mg L^−1	0	0
C(SO4^2−) actual value, mg L^−1	149.03	268.25
C(SO3^2−) actual value, mg L^−1	5657.31	11 314.62
C(SO3^2−) calculation value, mg L^−1	5829.40	11 658.79
C(SO3^2−) error, %	2.95	2.95
C(NO3^−) actual value, mg L^−1	189.80	370.04
C(NO3^−) calculation value, mg L^−1	200.79	396.95
C(NO3^−) error, %	5.47	6.78

---

As above mentioned, the catalyst supports have an obvious effect on surface properties. Therefore, it will be significant to investigate the effect of surface properties, such as surface area, surface acidity and points of zero charge values on the removal efficiency. The BET values in Table S1† show that the parameters of pore structures remain low among these different catalysts. It indicates that the parameters of pore structures have no influence on the removal efficiency.

Effect of surface acidity

NH₃-TPD profiles of the samples are shown in S4.† There are strong NH₃ desorption peaks over Fe–Si, but much weaker peaks are observed over Fe, Fe–Ti and Fe–Al. The number of acidic sites is obtained to be decreased as the following order: Fe–Si > Fe > Fe–Ti > Fe–Al. FTIR spectra of NH₃ adsorbed onto Fe, Fe–Al, Fe–Ti and Fe–Si are shown in Fig. 3. The band at 1600 cm⁻¹ has been assigned to the vibration modes of Lewis site coordinated NH₃, whereas the band at 1500 cm⁻¹ corresponds to a NH₄⁺ bonded to a Brønsted site. For Fe–Ti and Fe–Si, the amount of L acid sites increases apparently, whereas it does not change significantly for Fe and Fe–Al. It is clear that a main effect for Fe–Si and Fe–Ti is to create more Lewis acid sites on the surface. Experimental results show that the content of oxygen with Fe–Ti and Fe–Si is much higher than that of Fe and Fe–Al (as shown in the ESI, S5†). Combining with TPD and FTIR experiments, it can be concluded that the oxygen prefers to be generated on L acid sites. Pham et al. has reported that the H₂O₂ decompositions have two mechanisms.⁷ One is the ˙OH radical generation mechanism (the radical mechanism). The other is oxygen generation mechanism without the generation of ˙OH radicals (the non-radical mechanism). The decomposition mechanism of H₂O₂ maybe tends to the non-radical mechanism on Fe–Ti and Fe–Si due to the increase of the amount of Lewis acidic sites.

Fig. 3 FTIR spectra of Fe, Fe–Al, Fe–Ti and Fe–Si after adsorption of NH₃.

Surface information on Fe–Al and Fe–Ti was analyzed by XPS as shown in Fig. 4. The Fe peaks are assigned to oxidized Fe species.^12,13 The binding energies centered at about 709.6 and 723.4 eV may be assigned to Fe²⁺ in the spinal structure, and the binding energy of centered at about 711.4 and 725.3 eV can be ascribed to Fe³⁺ in the spinal structure.^14,15 The binding energy of centered about 713.6 eV is attributed to Fe³⁺ bonded with hydroxyl group (–OH).^13,16 The O peaks mainly centered at about 530 eV, as expected for the transition metal oxides. Another oxygen species centered at about 531 eV was assigned to O in –OH.^13,16 As shown in Fig. 4, more content of –OH is observed on the surface of Fe–Al. Results of NH₃-TPD experiments show that more number of acid sites is over Fe–Ti, which means the number of acid sites has direct relation with surface –OH. It has been reported that surface –OH could be acted as Brønsted acid.¹⁷ Moreover, it is worth noting that although the content of Fe²⁺ and –OH on Fe–Al is much lower than those on Fe–Ti, the content of Fe³⁺ bonded with –OH is much higher on Fe–Al. The NO_X removal efficiency of Fe–Al is much higher than that of Fe–Ti. It can be summarized that the variation of Fe valence and the number of B acidic sites affect the catalyst activity slightly, and the Fe³⁺ bonded with –OH plays the key role in the catalyst activity. In addition, more Fe(III) is detected on Fe–Ti (as shown Table S3 in the ESI†), which is also proved to contain more Lewis acid sites on the surface. Therefore, it can be speculated that the Fe(III) are most probably acted as Lewis acid sites.

Fig. 4 XPS spectra of fresh Fe–Al and Fe–Ti over the spectral regions of Fe 2p and O 1s.

Effect of points of zero charge

The different removal efficiencies of catalysts may arise from the difference of electronic properties because titania, alumina, iron oxides and silica exhibit different points of zero charge (pzc) values. The pzc values of these metal oxides were reported to be decreased as the following order: Al₂O₃ > Fe₂O₃ > TiO₂ > SiO₂.¹⁸ As shown in Fig. 2, the NO_X removal efficiency of Fe is only a little lower than that of Fe–Al, but the removal efficiency of Fe–Ti and Fe–Si is apparently lower, especially for Fe–Si. From XRD characterizations, it can be found that the predominant component in Fe and Fe–Al is hematite, and the characteristic peaks of hematite are clearly observed. However, those characteristic peaks of hematite in Fe–Ti and Fe–Si dramatically decrease, which means the main effect of Fe–Ti and Fe–Si on surface properties may be gradually changed to be carriers. Results of titration also indicate that the pzc values of these catalysts decrease as the following order: Fe–Al ≈ Fe > Fe–Ti > Fe–Si. The pzc of anatase (iep., 4) and silica (iep., < 4) is much lower than hematite (iep., 9.2) and alumina (iep., 9.6). At circumneutral pH values (pH = 5.5), it is expected that the surfaces of Fe and Fe–Al will be positively charged, whereas Fe–Ti and Fe–Si surfaces will be less positively charged or even negatively charged. The positive surface charge of Fe and Fe–Al will promote the adsorption of H₂O₂ on the surface of catalyst due to the strong nucleophilic ability of H₂O₂, which has four pairs of lone electron. The adsorption of H₂O₂ on the catalyst has been reported to be the rate-determining step in H₂O₂ decomposition.¹⁹ The similar removal efficiencies of Fe and Fe–Al are due to the same predominant component.

Effect of H₂O₂ flow

Since too much H₂O₂ had the explosion risk over 130 °C, the effect of H₂O₂ flow on NO_X and SO₂ removal efficiencies was further investigated. As shown in Fig. 5, increasing the H₂O₂ flow only has little effect on the SO₂ removal efficiency, but the effect is significant for the NO_X removal. As the H₂O₂ flow increases from 0.32 to 0.40 mL min⁻¹, the NO_X removal efficiency apparently increases. But further increase of H₂O₂ flow results in a constant or even a decline of NO_X removal efficiency. Fe–Ti and Fe–Si show more apparent decline in NO_X removal efficiency comparing with Fe–Al. It indicates that the H₂O₂ plays a key role in generating ˙OH radicals and removal of NO_X and SO₂. The increase of H₂O₂ flow can not only benefit the generation of ˙OH radicals, but also accelerate the annihilation of free radicals. At low H₂O₂ flow, the increase of H₂O₂ flow can result in the increase of ˙OH radicals due to the low annihilation of free radicals. However, at high H₂O₂ flow, too much ˙OH radicals will accelerate the annihilation of free radicals, thus inhibiting the ˙OH radical oxidation ability. Maybe the high pzc values and Fe³⁺ bonded with –OH inhibit the ˙OH radical annihilation, thus leading to the constant removal efficiency even though at high H₂O₂ flow. The high sulfite concentration was detected for the absorbent in the glass made ammonia-based washing tower with the ion chromatograph as shown in S7 (in the ESI, S7†). As we known, the sulfite is instability in the H₂O₂ solution. The high sulfite concentration indicates that most of H₂O₂ is decomposed under the catalysis of Fe based catalysts. Only little H₂O₂ can be injected into duct, thus sufficiently inhibiting the possibility of explosion. The NO_X removal efficiency is a little low in this process (about 74%). This may be caused by the low catalytic decomposition of H₂O₂ to ˙OH radicals. A study improving the catalytic activity of Fe based catalysts is in progress.

Fig. 5 Variations of NO_X and SO₂ removal efficiencies with H₂O₂ flow. Reaction conditions: 0.6 g catalyst, [NO_X] = 550 ppm, [SO₂] = 17 000 ppm, total flow = 240 mL min⁻¹, flow of free radical carrier gas = 40 mL min⁻¹, GHSV = 36 000 h⁻¹, simulated flue gas temperature = 130 °C.

Conclusions

Simultaneous removal of SO₂ and NO_X with H₂O₂ over Fe based catalysts were investigated in this paper. 100% SO₂ removal efficiency was obtained, and the maximum NO_X removal efficiency was about 74%. Effect of pzc values and surface acidity on NO_X removal efficiency was investigated. On Lewis acid sites, H₂O₂ decomposition tends to be the non-radical mechanism, and inhibit the generation of ˙OH radicals. The Fe³⁺ bonded with –OH plays the key role in the ˙OH radical generation. The valence of iron and the surface –OH have a only slight influence on the ˙OH radical generation. The increase of pzc values can benefit the adsorption of H₂O₂. Results of ion chromatograph indicate that most of H₂O₂ decomposes under the catalysis of Fe-based catalysts, and only NO₃⁻, SO₃²⁻ and SO₄²⁻ are produced in this process.

Acknowledgements

This work was finally supported by the National Natural Science Foundation of China (U1162119), The Assembly Foundation of the Industry and Information Ministry of the People's Republic of China 2012 (543), Research and Innovation Plan for Postgraduates of Jiangsu Province (CXZZ13_0215), Research Fund for Scientific Research Project of Environmental Protection Department of Jiangsu Province (2013003) and (201212), Industry-Academia Cooperation Innovation Fund Projects of Jiangsu Province (BY2012025) and Scientific Research Project of Environmental Protection Department of Jiangsu Province (201112).

References

1. R. Qi and R. T. Yang, Appl. Catal., B, 2003, 44, 217–225.
2. Y. Liu, J. Zhang, C. Sheng, Y. Zhang and L. Zhao, Chem. Eng. J., 2010, 162, 1006–1011.
3. M. Bai, Z. Zhang and M. D. Bai, Environ. Sci. Technol., 2012, 46, 10161–10168.
4. R. P. Dahiya, S. K. Mishra and A. Veefkind, IEEE Trans. Plasma Sci., 1993, 21, 346–348.
5. S. S. Lin and M. D. Gurol, Environ. Sci. Technol., 1998, 32, 1417–1423.
6. C. M. Lousada and M. Jonsson, J. Phys. Chem. C, 2010, 114, 11202–11208.
7. A. L. Pham, C. Lee, F. M. Doyle and D. L. Sedlak, Environ. Sci. Technol., 2009, 43, 8930–8935.
8. M. Hermanek, R. Zboril, I. Medrik, J. Pechousek and C. Gregor, J. Am. Chem. Soc., 2007, 129, 10929–10936.
9. V. R. Choudhary and S. D. Sansare, Ind. Eng. Chem. Res., 2007, 46(10), 3237–3242.
10. H. F. Schmidt, M. Meuris, P. W. Mertens, A. L. P. Rotondaro, M. M. Heyns, T. Q. Hurd and Z. Hatcher, Jpn. J. Appl. Phys., 1995, 34, 727–731.
11. J. N. Park, J. K. Shon, M. Jin, S. H. Hwang, G. O. Park, J. H. Boo, T. H. Han and J. M. Kim, Chem. Lett., 2010, 39, 493–495.
12. T. Herranz, S. Rojas, M. Ojeda, F. J. Perez-Alonso, P. Tefferos, K. Pirota and J. L. G. Fierro, Chem. Mater., 2006, 18, 2346–2375.
13. S. Yang, Y. Guo, N. Yan, Z. Qu, J. Xie, C. Yang and J. Jia, J. Hazard. Mater., 2011, 186, 508–515.
14. C. J. G. Peter and M. A. J. Somers, Appl. Surf. Sci., 1996, 100/101, 36–40.
15. X. Hu, J. C. Yu, J. Gong, Q. Li and G. Li, Adv. Mater., 2007, 19, 2324–2329.
16. S. Yang, Y. Guo, N. Yan, N. Yan, D. Wu, H. He, Z. Qu and J. Jia, Ind. Eng. Chem. Res., 2011, 50, 9650–9656.
17. T. W. Cynthia and D. C. Clarence, J. Phys. Chem., 1985, 89, 1569–1571.
18. M. Kosmulski, J. Colloid Interface Sci., 2002, 253, 77–87.
19. G. E. Marnellos, M. A. Efthimiadis and I. A. Vasalos, Appl. Catal., B, 2004, 48, 1–15.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c3ra46418k

---