CET Volume 86 DOI: 10.3303/CET2186067 Paper Received: 3 September 2020; Revised: 8 March 2021; Accepted: 6 May 2021 Please cite this article as: Flagiello D., Di Natale F., Erto A., Lancia A., 2021, Oxidative Scrubber for NOX Emission Control Using NaClO2 Aqueous Solutions, Chemical Engineering Transactions, 86, 397-402 DOI:10.3303/CET2186067 CHEMICAL ENGINEERING TRANSACTIONS VOL. 86, 2021 A publication of The Italian Association of Chemical Engineering Online at www.cetjournal.it Guest Editors: Sauro Pierucci, Jiří Jaromír Klemeš Copyright © 2021, AIDIC Servizi S.r.l. ISBN 978-88-95608-84-6; ISSN 2283-9216 Oxidative Scrubber for NOx Emission Control Using NaClO2 Aqueous Solutions Domenico Flagiello*, Francesco Di Natale, Alessandro Erto, Amedeo Lancia Department of Chemical, Materials and Production Engineering, University of Naples Federico II, P.le Tecchio, 80 - 80125 Naples, Italy domenico.flagiello@unina.it Oxidative scrubbing using different oxidizing chemicals (e.g. sodium chlorite, NaClO2) has been proposed since several years for gas emissions control. Although higher performances than conventional after-treatment process currently in use can be reached, these technologies have not yet found a consolidated application in the power generation and other industrial fields, due to excessive chemical consumption and the formation of some harmful contaminants in the wash water that must be removed with specific treatment systems. This work reports an experimental investigation of an oxidative wet scrubber for NOx removal from a simulated flue-gas containing 1030 ppmv of nitrogen oxides, using a scrubbing liquid doped with sodium chlorite (NaClO2) as oxidizing chemical. The experiments are performed in a packed column operating in counter- current flow with a constant gas velocity of 1.15 m/s at 60 °C and a liquid-to-gas ratio variable from 1.25 to 4.06 L/m3. The packed column with a DN 100 is equipped with 892 mm of Mellapak 250.X packing. The experimental removal efficiencies of NOx are evaluated in function of the liquid flow rate and NaClO2 loading. These data allow to determine the specific NaClO2 dosage required to achieve a fixed NOx removal. Finally, a predictive correlation for wet oxidative scrubbing is also developed. 1. Introduction Nowadays, the NOx emission control from flue-gases is still one of the major issues in the fields of energy generation, municipal waste incineration, internal combustion engines and industrial combustion processes. After-treatment de-NOx systems are the only option for existing plants and, despite the significant efforts made to improve design and operation of combustion systems aimed to reduce the NOx emissions in the flue-gases, they still remain a main option also for new installations. The Selective Catalytic Reduction (SCR) or thermal Selective Non-Catalytic Reduction (SNCR) units are the most used technologies for NOx emission control (Muzio et al., 2002; Van Caneghem et al., 2016). SCR is probably the most adopted after-treatment process and has been the subject of several technical and scientific studies aimed to improve conventional units and solve the typical drawbacks on catalyst operational, such as sintering due to high operating temperatures, sulphur and metal poisoning, catalysis competition effects and catalyst longevity (Gao et al., 2017). In spite of the improvements, these units suffer for the high capital investments (CAPEX) and operation and maintenance (O&M) costs, and when the flue-gas is cold or contains large amount of aerosols. In addition, their integration with desulphurization wet scrubbers, needed to assure a fair reduction of SOx compounds, is still not straightforward. In this paper, we propose a NOx removal system based on a wet oxidative scrubbing process using a chemical oxidant dissolved in aqueous solutions (Brogren et al., 1998). This technology potentially offers simpler and lighter equipment than SCR units, also determining much lower installation costs, if the oxidant consumption is optimized with a scrubber recirculation system and a specific water treatment plant is installed after scrubbing. In addition, the low operating temperature makes this unit very suitable as an end-of-pipe process for retrofitting of existing SCRs/SNCRs when further NOx reductions are mandatory. This allows to comply with the more stringent regulations recently set at European level (Directive (EU) 2016/2284), which require a reduction of NOx emissions of about 35 - 40% by 2030 and 60 - 70% after 2030. In this work, we report preliminary findings on NOx emission control by an oxidative scrubbing. In details, the oxidative scrubber operates with sodium chlorite (NaClO2), which is a very strong chemical oxidant 397 when dissolved in aqueous solutions. The experiments were performed in a pilot-scale scrubber having a diameter of 0.1 m and a height of 0.892 m, filled with Mellapak 250.X structured packing. The tests were performed with a simulated flue-gas stream with a constant gas flow rate of 32 m3/h containing 1030 ppmv of NOx at 60 °C (typical value of the end-of-pipe processes). The scrubbing solutions was an aqueous solution containing a concentration of NaClO2 ranging from 0 to 1% w/w, fed with liquid to gas ratios between 1.25 - 4.06 L/m3. 2. Absorption mechanism of NOx in NaClO2 aqueous solutions NO(g) is a gas almost insoluble in water and it accounts for more than 90% of NOx in a typical flue-gas, where the rest is mainly NO2(g). NOx absorption from a gaseous stream in water is characterized by a complex network of reactions, involving many compounds in both gas and liquid phase, and their interactions make this process much complicated (Hoftyzer and Kwanten, 1972). Indeed, there are many parallel and series reactions together with absorption and desorption phenomena of the species in solution, simultaneously (Figure 1). Figure 1: Reactive scheme for NOx absorption in water (Hoftyzer and Kwanten (1972)) NO(g) absorption is mainly related to both NO2(g) formation, which is about one order of magnitude more soluble in water than NO, and also N2O4(g) and N2O3(g) formation, both much more soluble in water than nitrogen dioxide (Hoftyzer and Kwanten, 1972; Jethani et al., 1992). The presence of these compounds promotes the HNO2(aq) and HNO3(aq) formation in water and influences the equilibria in the gas-phase. Due to its scarce solubility in water or into more alkaline solutions, NO(g) can desorbs and thus can be absorbed again or partially oxidized to NO2(g) in the gas-phase , producing HNO2(g), and partly reacts forming the two intermediates, N2O3(g) and N2O4(g), reiterating the reactive loop in liquid-film above described. The absorption of NOx in water is particularly influenced by the kinetic of NO(g) oxidation into NO2(g), which represents the slow step of the whole process (Miller, 1987; Thiemann, 2000). A rigorous equilibrium and kinetic model that describes the complex chemical and physical phenomena in the two phases for the absorption of NOx in water, containing the constants for phase and chemical equilibria and kinetic parameters, is reported in Flagiello (2020). In order to achieve a high removal efficiency, NO can be converted to NO2 in liquid-phase by using strong oxidants, such as sodium chlorite (NaClO2). Chlorite ions is mainly consumed to oxidize NO(aq) into NO2(aq) and then in nitrates, while ClO2 − is reduced to chlorides. Recent reports (Deshwal and Lee, 2009; Park et al., 2015) suggest that the oxidation reaction pathways with NaClO2 can be divided into two different reactive pathways, depending on whether the reaction occurs under acidic conditions or not. The mechanism of NOx absorption in NaClO2 solutions could be studied by considering only the oxidation reaction pathway with chlorite due to the poor solubility of NO(g) and NO2(g) in water and the very slow related reaction pathway (see Figure 1). The oxidation reactions for NOx absorption at pH > 7 (Basic Oxidation Mechanism, BOM) are reported below: ( ) 2 2( )2 2 2aq aqNO ClO NO Cl − −+ ↔ + (1) 2( ) 2 3 2 32 3 2aqNO H O H O NO NO + − −+ ↔ + + (2) 2 2 30.5 0.5NO ClO NO Cl − − − −+ ↔ + (3) 2( ) 2 2 3 34 6 4 4aqNO ClO H O H O NO Cl − + − −+ + ↔ + + (4) where the Eq. (4) is given by the sum of the Eqs. (2)-(3), and represents the overall reaction for NO2 oxidation. Standard Gibbs free energies of all reactions show that NOx is spontaneously absorbed and oxidized to HNO3(aq), as also confirmed in the studies of Chien et al. (2003). The addition and/or formation of acid (H3O +) 398 resulting from the BOM mechanism, could trigger a second oxidation pathway when the pH falls below 7 (Flagiello, 2020) and chlorite ions are converted into chlorine (Cl2(aq)) and chlorine dioxide (ClO2(aq)), which are stronger oxidants that are able to increase the oxidation rate of NOx. The reactions of formation of secondary oxidants are reported in Flagiello et al. (2020a), together with the equilibrium constants experimentally retrieved. The mechanism under acidic conditions (Acidic Oxidation Mechanism, AOM) is shown below. ( ) 2( ) 2 2( ) 35 2 3 5 2 2aq aq aqNO ClO H O NO H O Cl + −+ + ↔ + + (5) ( ) 2( ) 2 2( ) 33 2 2aq aq aqNO Cl H O NO H O Cl + −+ + ↔ + + (6) 2( ) 2( ) 2 3 35 9 6 5aq aqNO ClO H O H O NO Cl + − −+ + ↔ + + (7) 2( ) 2( ) 2 3 32 3 12 8 2 6aq aqNO Cl H O H O NO Cl + − −+ + ↔ + + (8) A rigorous equilibrium model, including the constants of oxidation reactions, is reported in Flagiello (2020). 3. Materials and Methods 3.1 Materials The simulated flue-gas was prepared by mixing NO at 2% v/v in N2 stored in high-pressure cylinders with compressed air at technical grade. The further mixing of O2 contents in air with NO stream allows obtaining NO2 in simulated flue-gas. The scrubbing liquid stream was prepared with a tap water at pH = 7.6 and its chemical composition, obtained by ionic chromatography method, is reported in Flagiello et al. (2018a). Sodium chlorite (NaClO2) was available at technical grade equal to 80% (w/w). The addition of sodium chlorite increased the pH value of the scrubbing solutions, which is equal to 8.96 and 9.40 for NaClO2 at 0.5% (w/w) and 1% (w/w), respectively. 3.2 Experimental set-up The flowsheet of the experimental set-up, inclusive of all the column equipment and measuring and analytical instruments, is shown in Figure 2. Figure 2: Flowsheet of the experimental set-up (Flagiello et al., 2020b) The experiments were performed in a Plexiglas column operating at 1 bar, with a DN 100 and a packing height of 0.892 m using Mellapak 250.X packing. Mellapak 250.X modules are made in Hastelloy C-22 alloy to prevent acid corrosion; other geometric and physical characteristics are reported in Flagiello et al. (2018b; 2020c). The feeding gas section was managed via digital flow meters by SMC Corporation (with accuracy of ±1NL/min) and its temperature was set at 60 °C using an electric gas heater provided by Megaris srl (total power of 1 kW), which was connected to a Proportional–Integral–Derivative (PID) controller for temperature 399 setting at the bottom inlet of the column. The column is also equipped with a gas pressure drop meter with a differential pressure gauge (FLUKE Corporation, 922 model with accuracy of ±0.1 mmH2O). Moreover, the gas temperature profile was measured with a digital thermometer (PCE T-390 model with accuracy ±0.1 °C) and the relative humidity at top and bottom was determined using a digital humidity controller (HOBO® onset UX100-23 model, with accuracy of ±0.1% of relative humidity). The scrubbing liquid was fed at the top of the column by a Grundfos Lenntech centrifugal pump (total power 0.75 kW) and controlled with a Cryotek Engineering flow meter (with accuracy of ±5L/h). The pH and temperature of the feeding liquid were measured with a digital pH-meter (PCE-228 model, with accuracy of ±0.01 of pH and ±0.01 °C for temperature). The liquid was atomized in the column by a PNR® full cone nozzle (DAM 1212 B31 model) with a complete opening of the liquid jet of 45°, and a 90 mm height plastic foam demister was put at 15 mm from the nozzle at the top of the column to block the entrained liquid drops. 3.3 Methodology of running tests and instrumental analytics The experimental runs were carried out by feeding the simulated flue-gas stream to the column with a constant flow rate of 32 m3/h corresponding to 1.15 m/s, which is the typical gas velocity in a de-SOx scrubber (Flagiello et al., 2019). The gas temperature and its relative humidity were set to 60 °C and 10 - 15%, respectively. The NOx fed composition was set to 1030 ppmv and checked by a dedicated gas analyser (Eco Physics CLD 62) using the pipe-line before the scrubber. The scrubbing liquid stream was sent in counter- current flow with the gas at different flow rates, from 40 to 130 L/h. The liquid temperature was set at 25 °C, while the NaClO2 dosages ranged from 0 to 1% (w/w). The NOx gas concentration at the outlet of the column was monitored by the Eco Physics CLD 62 gas analyzer (limit of device is 100 up to 10000 ppmv), which allows measuring both NO and total NOx concentration. A gas sampling system was installed upstream to the gas analyzer, consisting of a KNF diaphragm pump (NMP 830 HP model), a Key Instruments flow meter (2500 Series, up to 1 L/min) and a Bühler Technologies gas quencher (TC-Standard Series). The experimental NOx removal efficiency (ηNOx) was calculated by comparing the input and output NOx concentration, as by Eq. (9). , , , x x in x out NO x in NO NO NO η − = (9) The wash water was collected at the bottom of the column and sent to a sampling point for further analysis, i.e. temperature and pH value by digital pH-meter (PCE-228 model). 4. Results and Discussions Figure 3 shows the experimental removal efficiencies of total NOx (A) and the final wash water pH (B) as a function of the liquid-to-gas volumetric ratio (L/G), varied between 1.25 and 4.06 L/m3. The results are shown parametrically with NaClO2 loading (0 - 1% w/w). Figure 3: Experimental results of total NOx removal efficiency (A) and wash water pH (B) as a function of the Liquid-to-Gas ratio (L/G) and parametric with NaClO2 loading in the scrubbing solution (from 0 to 1% w/w) Figure 3A shows that the NOx removal efficiency in tap water (i.e. with 0% w/w NaClO2 loading) is very low, up to 2%, consistently with the theory reported in the open literature and experiments with seawater by Flagiello et al. (2021). This was due to the low solubility of NO and NO2 in water and to the very slow hydrolysis mechanisms (see Figure 1). On the contrary, the NOx removal efficiency increased up to 35% by increasing 400 the NaClO2 content. An appreciable influence of the liquid flow rate (or L/G ratios) was observed under these operating conditions, and the effects resulted more pronounced for NaClO2 at 1% w/w. Figure 3B shows that the pH in the wash water solutions were negligibly affected by the L/G ratio. Besides, both the tap water and the NaClO2 solutions exhibited a final pH very close to the initial values (7.60, 8.94 and 9.40). While for the tap water no absorption takes place and, thus, the preservation of pH is expected, for NaClO2 solutions, this observation confirms that the reaction between chlorite and NO and NO2 in the aqueous phase occurred without altering the solution pH, according to the BOM mechanism reported in Eqs. (1)-(4). The results showed in Figures 3A-C were rearranged in Figure 4 to find a dependency between NOx removal efficiency and the operating dosage (dop) defined as the ratio between the molar dosage of ClO2 − in the scrubbing liquid and the NOx in the fed flue-gas. A predictive equation was developed and showed in Figure 4. Figure 4: Experimental and modelling results of NOx removal efficiency as a function of the operating dosage Figure 4 shows that the NOx removal efficiency increased with the operating dosage in the scrubbing solution. The data confirmed that NOx oxidation was not strictly dependent on liquid flow rate but rather on the amount of NaClO2 used as compared to the NOx in the gas fed. Figure 4 also shows that NOx removal efficiency can be adequately described by a power law function in the tested range of NaClO2 dosages and, within certain validity limits, could allow to predict the performance of the process for higher operating dosages. The maximum experimental removal achieved (35%) required a molar quantity of chlorite ions about 12 times greater than the NOx one. As expected, to achieve higher NOx removal values, the modelling equation predicts that an increasing excess of sodium chlorite must be added. For a further reduction of emission limits as planned before and after 2030, the performance required for an existing de-NOx may be achieved using oxidative scrubbers as an end-of-pipe treatment. For example, considering that the further NOx reduction must be equivalent to 60% of NOx removal efficiency in the scrubber, a NaClO2 dosage about 50 times the NOx molar fed is required. The large excess used could be a drawback for this operation, but a recirculation of a fraction of the wash water still containing the excess oxidant could make the process more attractive, reducing the costs. 5. Conclusions This work aimed to evaluate the oxidative performances of sodium chlorite towards NOx, to enhance the absorption of this compound, which is poorly soluble in water. The experiments were performed in a packed column operating with a model flue-gas at 1.15 m/s and 60 °C containing 1030 ppmv of NOx and a scrubbing solution containing NaClO2, with a dosage from 0 to 1% (w/w), at 25 °C. The experimental results showed that for an oxidative scrubber the removal efficiencies increased with the addition of NaClO2 and by increasing the liquid flow rate. A maximum removal efficiency (35%) was observed for 1% (w/w) NaClO2 loading and a liquid-to-gas ratio equal to 4.06 L/m 3, while a very lower efficiency (about 2%) using a tap water was obtained. The results also allowed to determine the main NOx oxidation pathway, which resulted to be the mechanism under basic conditions (BOM, see Eqs. (1)-(4)); indeed, the wash water pH resulted to be higher than 7 in all the investigated conditions. The nitric acid produced during oxidation was unable to reduce pH due to the buffering effect of the tap water, and the second oxidation pathway under acidic conditions (AOM) did not occur (Eqs. (5)-(8)). The operating molar dosage of NaClO2 that allowed to reach the maximum removal efficiency (35%) was equal to about 12 times the moles of NOx in the gas fed. These results suggested that higher removals may be achieved by further increasing the NaClO2 loading in the scrubbing liquid. The experimental data also 401 confirmed that the NOx removal efficiency can be expressed by a power law function with an exponent equal to 0.431 of the operating dosage, dop. This process could be suitable in the retrofitting of existing SNCR plants as an end-of-pipe process, where further efforts in terms of NOx removal are required to comply with the more stringent recent regulations that will be in force with Directive (EU) 2016/2284. New generation of SCRs could also benefit for the integration of an oxidative scrubber, suitable for the abatement of a certain amount of NOx, reducing CAPEX and O&M costs, and size of plant, as well as urea consumption and ammonia slip. Although a large excess of NaClO2 was necessary to achieve high levels of efficiency, as predicted by the retrieved model, the integration of a closed-loop scrubber system is recommended to optimize the oxidant consumption, hence reducing the costs of the chemicals. Further efforts are needed to investigate the oxidative scrubber operating in a closed-loop system to optimize the consumption of NaClO2 and the scrubber performance. Future work will be also focused on the integration of a specific wash water treatment system after the oxidative scrubbing process and on the presence of SO2 in the simulated flue-gas or an artificial acidification of the oxidizing liquid in light of a potential improvement in the oxidative performances, as suggested by the literature. Acknowledgments The Authors warmly acknowledge Eng. Maria Montieri for the support in the experimentation. References Brogren C., Karlsson H. T., Bjerle I., 1998, Absorption of NO in an aqueous solution of NaClO2, Chem. Eng. Technol., 21:1, 61-70. Chien T. W., Chu H., Hsueh H. T., 2003, Kinetic study on absorption of SO2 and NOx with acidic NaClO2 solutions using the spraying column, J. Environ. Eng., 129:11, 967–974. Deshwal B. R. and Lee H. K., 2009, Mass transfer in the absorption of SO2 and NOx using aqueous euchlorine scrubbing solution, Int. J. Environ. Sci., 21:2, 155–161. Flagiello D., Erto A., Lancia A., Di Natale F., 2018a, Experimental and modelling analysis of seawater scrubbers for sulphur dioxide removal from flue-gas, Fuel, 214, 254–263. Flagiello D., Di Natale F., Carotenuto C., Erto A., Lancia A., 2018b, Seawater Desulphurization of Simulated Flue Gas in Spray and Packed Columns: an Experimental and Modelling Comparison, CET, 69, 799–804. Flagiello D., Parisi A., Lancia A., Carotenuto C., Erto A., Di Natale F., 2019, Seawater desulphurization scrubbing in spray and packed columns for a 4.35 MW marine diesel engine, ChERD, 148, 56–67. Flagiello D., 2020, Enhanced seawater scrubbing for flue-gas cleaning, PhD Thesis, University of Naples Federico II, Naples, Italy. Flagiello D., Di Natale F., Erto A., Lancia A., 2020a, Wet oxidation scrubbing (WOS) for flue-gas desulphurization using sodium chlorite seawater solutions, Fuel, 277, 118055. Flagiello D., Erto A., Lancia A., Di Natale F., 2020b, Dataset of wet desulphurization scrubbing in a column packed with Mellapak 250.X, Data in Brief, 33, 106383. Flagiello D., Di Natale F., Lancia A., Erto A., 2020c, Characterization of mass transfer coefficients and pressure drops for packed towers with Mellapak 250.X, ChERD, 161, 340–356. Flagiello D., Di Natale F., Lancia A., Salo,K., 2021, Effect of Seawater Alkalinity on the Performances of a Marine Diesel Engine Desulphurization Scrubber, CET, 86. Gao F. et al., 2017, A review on selective catalytic reduction of NOx by NH3 over Mn–based catalysts at low temperatures: catalysts, mechanisms, kinetics and DFT calculations. Catalysts, 7:7, 199. Hoftyzer P. J., Kwanten F. J. G., 1972, G. Nonhebel (Ed): Gas Purification Processes for Air Pollution Control, Butterworths, London, UK. Jethani K. R., Suchak N. J., Joshi J. B., 1992, Modeling and Simulation of a Spray Column for NOx Absorption, Comput. Chem. Eng., 16:1, 11–25. Miller D. N., 1987, Mass transfer in nitric acid absorption, AIChE journal, 33:8, 1351–1358. Muzio L. J., Quartucy G. C., Cichanowiczy J. E., 2002, Overview and status of post-combustion NOx control: SNCR, SCR and hybrid technologies, Int. J. Environ. Pollut., 17:1-2, 4-30. Park H. W., Choi S., Park D. W., 2015, Simultaneous treatment of NO and SO2 with aqueous NaClO2 solution in a wet scrubber combined with a plasma electrostatic precipitator, J. Hazard. Mater., 285, 117–126. Thiemann M., Scheibler E., Wiegand K. W., 2000, Nitric acid, nitrous acid, and nitrogen oxides, Chapter In: Ullmann's encyclopedia of industrial chemistry, Vol 26, Wiley-VCH, Weinheim, Germany, 177–225. Van Caneghem J. et al., 2016, NOx reduction in waste incinerators by selective catalytic reduction (SCR) instead of selective non catalytic reduction (SNCR) compared from a life cycle perspective: a case study, J. Clean. Prod., 112, 4452-4460. 402