Nitrogen oxide removal using seawater electrolysis in an undivided cell for ocean-going vessels

Abstract

As massive nitrogen oxide (NO_x) emissions from marine slow speed diesel engines have caused serious health and environmental problems, NO_x removal using electro-generated chlorine under seawater electrolysis was studied in a lab-scale scrubbing reactor. Electro-generated chlorine was prepared by seawater electrolysis in an undivided cell. Results showed that active chlorine concentration increased linearly with the increase of current density and electrolysis time. Energy consumption decreased from 83 to 6.1 kW h per (kg [Cl₂]) with the salinity varying from 3.2 to 37.8 ppt after 5 min of electrolysis. Onsite generation of active chlorine using concentrated seawater from desalination plants operated on a ship was suggested to be a cost-effective solution with less energy consumption. Then, effects of dilution ratio (seawater/electrolyte), inlet NO and SO₂ concentration, and initial pH value of scrubbing solution on NO_x removal efficiency were further investigated. NO_x removal efficiency of electro-generated chlorine at pH value of 7 decreased from 98.9% to 20.2% with the dilution ratio increasing from 2.4 to 96. The NO absorption rate by HOCl was proved to be higher than that by OCl⁻. NO absorption rate increased linearly as inlet NO concentration increased. When SO₂ concentration increased from 207 to 814 ppm, NO_x removal efficiency decreased slightly, but SO₂ removal efficiency was almost kept at 100%. The possible reaction mechanism and pathways were discussed by evaluating the pH-dependant NO_x removal efficiency using electro-generated chlorine. The proposed method of wet scrubbing using electro-generated chlorine was demonstrated to be a potential after treatment strategy to control NO_x emissions from large marine diesel engines.

---

1. Introduction

With over 90% of world trade carried by sea, international shipping has caused an increasingly serious impact on air pollution over the past few decades.^1,2 Worldwide, from 2007 to 2012, shipping is estimated to produce approximately 18.6 million tons of NO_x annually, which accounts for approximately 15% of global NO_x emissions from anthropogenic sources.^2,3 NO_x plays an important role in the formation of acid rain, photochemical smog, aerosol warming, and toxic haze, which influence human lives and activities at regional scales.⁴ Currently, the International Maritime Organization (IMO) and some environmental protection agencies from developed countries implement more stringent regulations⁵ to limit NO_x emissions from shipping.

Various technologies have been proposed to meet these regulations on ship NO_x emissions,^6,7 including exhaust gas recirculation (EGR), selective catalytic reduction (SCR), non-thermal plasmas (NTP), electrochemical catalysis, wet scrubbing. Among these technologies, NH₃ SCR is the best-known technology, which is currently commercial available to achieve compliance with Tier III NO_x standards for marine engines.⁸ Until now, SCR systems have been successfully applied to four-stroke medium speed engines. However, only a few slow speed crosshead engines have been equipped with SCR. This is mainly because emissions of slow speed diesel engines burning low-quality fuels contain a high concentration of sulfur oxide (SO_x) and a low exhaust temperature (<300 °C),⁹ which deteriorate the catalyst performance in SCR systems. Therefore, wet scrubbing method, which uses chemical reagents to absorb NO_x gas via gas–liquid reactions, becomes a potentially attractive technique. Moreover, wet scrubbing technique can be coupled with the existing desulfurization technique, so that the investment cost for removing multiple pollutants can be saved.

More than 90% of ship NO_x emissions is in the form of nitric oxide (NO). For wet scrubbing systems, chemical additives are usually added into the scrubbing liquid to first convert relatively inert NO into more soluble NO₂, which is then removed by alkaline absorbents. A number of chemical oxidants¹⁰ have been investigated to determine their effectiveness in NO_x removal, including KMnO₄, H₂O₂, NaOCl, NaClO₂, O₃. However, the storage space required for most of the chemical oxidants is expected to be extremely large. This is mainly attributable to two reasons: (1) two-stroke slow speed diesel engines emit a high concentration of NO_x (typically 10 g kW⁻¹ h⁻¹), and (2) the replenishment interval of ocean-going vessels can be up to three months. As the engine room space for a ship is limited, the storage of abundant chemical oxidants will probably reduce the transport capacity and increase the potential risk of safety for the crew. Therefore, the interest in NaOCl wet scrubbing technique has increased in recent years because of its onsite production character by an electrochemical process.

Elena Ghibaudi et al.¹¹ studied the reaction between NO and NaOCl aqueous solution over a pH range of 6–12. They pointed that NO absorption rate by HOCl was much faster than that by OCl⁻. Luke Chen et al.¹² investigated NO absorption using NaOCl reagent and developed a mass transfer model for a packed-bed scrubbing system. Ruitang Guo et al.¹³ analyzed NO absorption kinetics with weak acidic NaOCl solution in a stirred tank reactor. Results showed that the absorption reaction was first-order with respect to both NO and NaOCl. Monoj Kumar Mondal and Chelluboyana Vaishnava Raghunath^14,15 studied optimal conditions for simultaneous absorption of NO and SO₂ into NaOCl freshwater solution. They found that NO removal efficiency could be influenced by several parameters, such as absorbent concentration, initial NO and SO₂ concentration, initial pH value of NaOCl solution.

Although many experiments have been conducted with NaOCl wet scrubbing technique, only a few studies have been done on NO removal using NaOCl solution generated under seawater electrolysis. Sukheon An and Osami Nishida^16,17 proposed a two-stage wet scrubbing system to control exhaust emissions from large marine diesel engines. They used a divided electrochemical cell to produce both acidic seawater and alkaline seawater, which are applied to oxidize NO to NO₂ and absorb NO₂ and CO₂, respectively. Nevertheless, a high concentration of total dissolved solids in seawater would accelerate the fouling of diaphragm in the divided cell. Thus, a short lifetime of ionic membrane could add extra operational cost of this method. In addition, Tae-Woo Kim¹⁸ investigated NO oxidation characteristic of seawater treated by undivided electrolysis in a bubbling reactor. Although some operational conditions (e.g., active chlorine concentration and temperature) were discussed, energy consumption and reaction mechanism of the NO absorption process remained unclear.

Therefore, the main objective of this study is to (1) determine effective operational conditions for onsite generation of active chlorine onboard with less energy consumption; (2) investigate NO_x removal performance of electro-generated chlorine. Effects of current density and seawater salinity on the production of active chlorine and energy consumption were studied in an undivided cell. After that, some parameters, such as dilution ratio, inlet NO and SO₂ concentration, and initial pH value of scrubbing liquid were investigated on NO_x removal efficiency. The possible reaction mechanism and pathways between NO and active chlorine species were discussed.

2. Materials and methods

2.1 Experiment apparatus and materials

A schematic diagram of the experimental apparatus is shown in Fig. 1. The experimental apparatus was composed of a simulated flue gas system, a spray scrubber, and a flue gas analysis system.

Fig. 1 Schematic diagram of experimental apparatus.

The flue gas of a diesel engine was simulated by mixing different gas species: NO (9.8% NO with 90.2% N₂ as balance gas), SO₂ (9.9% SO₂ with 90.1% N₂ as balance gas) and N₂ (99.9%). The flow rate of each material gas was regulated by a mass flow controller (MFC, D07-19B, Sevenstar Electronics, China). The NO and SO₂ concentration of the simulated flue gas were adjusted by the proportional ratio. All pipes, valves, regulators, and fittings were made up of stainless steel 316 grade or Teflon.

The spray scrubber (height 30 cm; internal diameter 6 cm) was a custom-made lucite spray column. The spray nozzle (B1/4TT-SS+TG-SS0.4, Spraying System Co., America) orifice was 0.04 cm. The feed pump was a peristaltic pump (YZ1115, Longer Precision Pump, China). The scrubbing liquid in the feeding tank (1.3 L) was mixed by a magnetic stirrer (750 rpm). In order to reduce the interference of photochemical decomposition by sunlight, the outside of the feeding tank was covered with aluminum foil.

The flue gas analysis system consisted of an electronic condenser and a gas analyzer (MGA 5, MRU, Germany). The MGA 5 gas analyzer was an infrared multigas analyzer with a detection range up to 2000 ppm at a resolution of 1 ppm, which was used for online measurement of inlet and outlet concentrations of NO, NO₂, NO_x, and SO₂.

The artificial seawater (salinity of 23.3 ppt) was prepared according to the ASTM D1141 standard as described previously.¹⁹ All chemicals used in our experiments were analytical reagent grade. The deionized water (15 MΩ cm) was prepared in a two-stage ELGA PURELAB Option R15 purification system.

The undivided cell for seawater electrolysis was a cylindrical lucite tank (2.2 L) with two plate electrodes (10 × 10 × 0.2 cm, Baoji Ruicheng Titanium Industry Co., Ltd., China). The anode material and cathode material were IrO₂–RuO₂/Ti and Ti, respectively. They were placed vertically in parallel with an inter-electrode gap of 2 cm. The electrolysis was performed using a DC power supply (GuDeng Electric GD-80V-25A, China) with a constant current control mode.

2.2 Experimental procedure

First, onsite generation performance of active chlorine by seawater electrolysis was investigated in the undivided cell. Before each operation, the cell was sonicated to completely desorb any remaining materials. During each run, the electrolyte was mixed by a magnetic stirrer, and a sample (5 mL) was withdrawn for measuring TOS concentration at an interval of 5 min. Then these samples were diluted when necessary and stored in a 4 °C water bath. Active chlorine concentrations of samples were determined by N,N-diethyl-p-phenylenediamine (DPD) colorimetric method (ASTM 4500 G), using a spectrophotometer (Shimadzu UV1800, Japan) at 515 nm.

Then, further experiments were performed to study NO_x removal characteristics by electro-generated chlorine in the custom-made spray scrubber. The active chlorine solution (2400 mg L⁻¹ [Cl₂]) was prepared by artificial seawater electrolysis under conditions of seawater salinity 23.3 ppt and current density 150 mA cm⁻². During each set of scrubbing experiments, the active chlorine solution was stored under seal in 20 °C water bath shielded from light, so that the decay of active chlorine was less than 2% in 10 h. After that, the scrubbing liquid was prepared by freshly mixing the electro-generated chlorine with artificial seawater. The initial pH value of scrubbing liquid was adjusted to the desired value by adding HCl solution (1 mol L⁻¹) or NaOH solution (1 mol L⁻¹). The pH value and oxidation–reduction potential (ORP) of the scrubbing liquid were measured by a pH/ORP meter (Mettler-Toledo S210, Switzerland).

As shown in Fig. 1, during each run, the simulated flue gas went through the main line (V1 open, V2 closed). When the inlet gas concentrations were stable, the scrubbing liquid was introduced into the main scrubber. The gas absorption reactions occurred at room temperature (25 °C). The outlet concentrations of NO, NO₂, NO_x and SO₂ were measured by the gas analyzer at an interval of 10 s during each run. The scrubbing liquid was supplied without recirculation, and the effluent liquid was collected in the bottom tank. After scrubbing, the simulated flue gas was discharged into atmosphere. Under all experiments, liquid flow rate and gas flow rate were set to 0.25 L min⁻¹ and 1.25 L min⁻¹, respectively.

2.3 Data processing

When the simulated flue gas reacted with the electro-generated chlorine in the counterflow scrubber, NO_x and SO₂ were removed. The removal efficiency (η_i, %) was defined aswhere i stands for NO, NO_x or SO₂; C_i,inlet is the inlet gas i concentration (ppm); C_i,outlet is the outlet gas i concentration (ppm). The NO_x is the sum of the NO and NO₂. The outlet concentrations of NO, NO₂ and SO₂ were obtained by taking averages after 2 min.

For the kinetic calculation, NO absorption rate (N_NO, mol (m⁻² s⁻¹)) was calculated by

where Q_g is the gas flow rate (L min⁻¹); A_i is the gas–liquid interfacial mass transfer area (m²); M_NO is the molecular weight of NO (M_NO = 30 g mol⁻¹).

3. Results and discussion

3.1 Production of active chlorine

Although seawater electrolysis is well known as an onsite active chlorine generation technology, its high energy consumption limits the application in marine equipment industry. To determine the feasibility of onsite generation of active chlorine on the ship with less energy consumption, we investigated effects of current density and seawater salinity on the production of active chlorine and energy consumption in an undivided cell.

3.1.1 Current density. Fig. 2 shows the effect of current density on the production of active chlorine as a function of electrolysis time (main graph) and the slope of active chlorine regression line as a function of current density (inset graph). Experiments were carried out in an undivided cell under condition of seawater salinity of 23.3 ppt. At all current densities, the active chlorine concentration increased linearly with increase of electrolysis time. The linear relationship between the active chlorine formation rate (i.e. the slope of active chlorine regression line) and the current density is presented in the inset graph in Fig. 2, which was similar to the results reported by Byung Soo Oh.²⁰ According to the Faraday's law, the active chlorine formation rate (mg (L⁻¹ min⁻¹) [Cl₂]) can be expressed aswhere ACC is the active chlorine concentration (mg L⁻¹ [Cl₂]); t is the electrolysis time (s); M_Cl2 is the molecular weight of chlorine (M = 71 g mol⁻¹); A_e is the area of the electrode (A_e = 100 cm²); n is the number of electrons transferred in the reaction at the anode (n = 2); F is the Faraday constant (F = 96 500 C mol⁻¹ electrons); V is the volume of the electrolyte solution (V = 2.2 L); CE is the current efficiency of the anode (%); i is the current density (mA cm⁻²).

Fig. 2 Active chlorine concentration as a function of electrolysis time at different current densities (main graph), and slope of active chlorine regression line as a function of current density (inset graph). Experimental condition: seawater salinity 23.3 ppt.

As the CE of electrolysis of highly concentrated electrolytes was almost constant under different current densities,²¹ the eqn (3) proved the linear relationship, which is shown in the inset graph in Fig. 2. This linear model implies that quantitative prediction of active chlorine is possible during seawater electrolysis using constant current mode. Moreover, this simple linear relationship also contributed to the easy control of seawater electrolysis cells onboard.

3.1.2 Seawater salinity. Considering that the salinity of seawater drawn through the sea chest of a ship varies from place to place, the effect of seawater salinity on active chlorine concentration and energy consumption are shown in Fig. 3 and 4, respectively. As seen in Fig. 3, active chlorine increased logarithmically as electrolysis time increased at all salinities, which was similar to those reported by Husam D. Al-Hamaiedeh.²² However, the active chlorine curves at salinities of no more than 7.2 ppt and no less than 13.2 ppt diverged after 20 min, which might be due to the temperature variation of electrolyte. When the seawater salinity was no more than 7.2 ppt, the resistive heating caused by the high cell resistance resulted in a sharp rise in the temperature of electrolyte. This rise in temperature promoted the conversion of active chlorine to chlorate and accelerated the dissolved chlorine to escape into the atmosphere, which resulted in the drop of the active chlorine formation rate after 20 min (Fig. 3). In contrast, the resistive heating caused by the cell resistance at salinity of no less than 13.2 ppt only resulted in a limited rise of the temperature of electrolyte. Consequently, the active chlorine formation rate at salinity of no less than 13.2 ppt decreased slightly with the electrolysis time varying from 20 to 30 min.

Fig. 3 Active chlorine concentration as a function of electrolysis time at different salinities. Experimental condition: current density 150 mA cm⁻².

Fig. 4 Energy consumption as a function of electrolysis time at different salinities. Experimental condition: current density 150 mA cm⁻².

As an energy-intensive process, the energy consumption of production of active chlorine was calculated as a function of seawater salinity at different electrolysis time under condition of current density of 150 mA cm⁻² (Fig. 4). The energy consumption decreased rapidly with an increase in the salinity, which was consistent with results reported by Hilary Nath et al.²³ The power type energy consumption profile was almost irrespective of electrolysis time.

The ohmic resistance of the electrolyte (R, Ω) is defined as

where k is the conductivity of the electrolyte (1/(Ω cm)); B is the conversion coefficient from salinity to conductivity (B ≈ 1560 (ref. 24)); l is the distance of the inter-electrode gap (cm); S is the salinity of the electrolyte (ppt).

According to the literature,²⁵ the applied cell voltage (U, V) can be given bellow

U = U_re + U_e + U_R + U_t (5)

where U_re is the thermodynamic (equilibrium) potential difference for the given electrode reactions (V); U_e is the sum of the anodic and cathodic overpotentials (V); U_R is the ohmic drop of the electrolyte (V); U_t is the drift of ΔU with electrolysis time due to degradation of the electrode performance, for fresh electrodes ΔU = 0 V.

For the specific undivided electrolysis cell in our experiments, we can simplify the eqn (5) into the following eqn (6), in which the applied cell voltage keeps a power function with the salinity of electrolyte.

where U′ and k′ are constant,

U′ = U_re + U_e, k′ = (i × l)/(1000 × B).

For seawater electrolysis process, heat dissipation at electrodes caused extra thermal energy loss, which resulted in a high energy consumption. According to the model proposed by Øystein Ulleberg,²⁶ thermal energy generated due to the energy inefficiency was linear to the difference between the applied cell voltage and the theoretical thermoneutral voltage. Therefore, from energy balance perspective, the energy consumption was expected to have a power function with the salinity, which was in agreement with nonlinear trends shown in Fig. 4.

Considering the actual situation of seawater on a ship, concentrated seawater can be obtained from desalination plants (e.g. vacuum boiling evaporator) operated on the ship, which is usually 1.5 to 2 times higher than the salinity of ambient seawater.²⁷ Therefore, onsite generation of active chlorine using concentrated seawater electrolysis was demonstrated to be a cost-effective solution with less energy consumption for ocean-going vessels.

3.2 NO_x removal by electro-generated chlorine

3.2.1 Dilution ratio of seawater to electrolyte. A set of experiments was performed to assess NO_x removal performance by electro-generated chlorine in a lab-scale scrubber. The effect of the dilution ratio of seawater to electrolyte on removal efficiencies is shown in Fig. 5.

Fig. 5 Removal efficiencies as a function of dilution ratio (seawater/electrolyte). Experimental conditions: pH 7 (solid points), uncontrolled pH (open points), inlet NO concentration 1000 ppm.

When the initial pH value of scrubbing solution was uncontrolled (above 8), the NO_x removal efficiency dropped from 62.2% to 6.57% with the dilution ratio varying from 1 to 32, which was due to the reduction of the electro-generated chlorine.¹⁵ By comparison, when the initial pH value of scrubbing solution was adjusted to 7, the NO_x removal efficiency by the electro-generated chlorine solution was more efficient than that without pH control at all range of the dilution ratio. Thus, the NO absorption rate by the active chlorine at pH value of 7 was higher than that at pH value of above 8.2. This result might suggest that HOCl was the more active than OCl⁻ in terms of NO absorption, even though its equilibrium concentration was much lower than the concentration of OCl⁻ in alkaline solution.

When the pH value of the electro-generated chlorine solution is above 7, the hypochlorite exists in equilibrium with the hypochlorous acid (i.e. dissociation of hypochlorous acid), depending on the pH value.²⁸

HOCl ↔ H⁺ + OCl⁻, pK_a = 7.54 (25 °C) (7)

Cl⁻ + 2OH⁻ → ClO⁻ + H₂O + 2e⁻, E⁰ = −0.89 V (8)

Cl⁻ + H₂O → HOCl + H⁺ + 2e⁻, E⁰ = −1.482 V (9)

According to the literature,²⁹ the NO removal efficiency was described as a function of the standard redox potentials of half reactions in which the reagents take part. As the standard oxidation potential of HOCl (E⁰ = −1.482 V) was lower than that of OCl⁻ (E⁰ = −0.89 V), NO removal efficiency was expected to decrease with the pH value of above 7. Moreover, these results might also be explained by the ORP of electro-generated chlorine solution as a function of pH (data not shown). The ORP was observed to decrease with the increasing pH value, which was consistent with the results reported by Cheryl N. James et al.³⁰ According to the dissociation of hypochlorous acid (reaction (7)), the concentration of HOCl at pH value of 7 is higher than that under pH value of above 8. Thus, the ORP of active chlorine solution decreased with the increasing pH value in weak alkaline solution, which eventually resulted in a reduction of NO_x removal efficiency.

3.2.2 Inlet NO concentration. Fig. 6 displays the effect of inlet NO concentration on NO and NO_x removal efficiencies. The NO removal efficiency increased from 14.2% to 75.9% with inlet NO concentration varying from 250 to 1250 ppm. Likewise, the NO_x removal efficiency increased from 10.2% to 59% with inlet NO concentration varying from 250 to 1250 ppm. In addition, as the gas and liquid flow rate were unchanged under all experiments, the gas–liquid interfacial mass transfer area of the scrubber was kept constant. According to the eqn (2), NO absorption rate was dependent on the difference between C_NO,inlet and C_NO,outlet. As shown in Fig. 7, this difference increased linearly as inlet NO concentration increased. Therefore, the NO absorption rate increased linearly as inlet NO concentration increased, which was in accordance with the NO absorption model proposed by Yangxian Liu et al.³¹ According to the two-film theory, an increase in inlet NO concentration would increase mass transfer driving force of NO gas in gas–liquid two phase and reaction rate of electro-generated chlorine with NO, which eventually improved NO absorption rate.

Fig. 6 Removal efficiencies as a function of inlet NO concentration. Experimental conditions: dilution ratio (seawater/electrolyte) 7, pH 7.

Fig. 7 Difference between C_NO,inlet and C_NO,outlet as a function of inlet NO concentration. Experimental conditions: dilution ratio (seawater/electrolyte) 7, pH 7.

3.2.3 Inlet SO₂ concentration. In general, SO₂ is one kind of harmful gaseous species from exhaust emissions of marine diesel engines. This acid gas also has a potential to affect NO removal. Therefore, experiments were performed to investigate SO₂ concentration on NO, NO_x and SO₂ removal efficiencies, and results are shown in Fig. 8. The SO₂ removal efficiency was almost kept at 100% under all SO₂ concentrations, which might suggest that SO₂ removal was controlled by liquid mass transfer. The NO and NO_x removal efficiencies decreased slightly as SO₂ concentration increased from 207 to 814 ppm, which might be resulted from a competition reaction between SO₂ and NO. These results implied that SO_x and NO_x could be simultaneously removed by sufficient amount of electro-generated chlorine. Therefore, the proposed method could have the potential to meet requirements of emission control regulations for ocean-going vessels using low-quality marine fuel oil.

Fig. 8 Removal efficiencies as a function of inlet SO₂ concentration. Experimental conditions: dilution ratio (seawater/electrolyte) 7, pH 7, inlet NO concentration 1000 ppm.

3.3 Reaction mechanism

To determine the reaction mechanism between NO_x and active chlorine solution, experiments were carried out to investigate the effect of initial pH value of scrubbing solution on the NO and NO_x removal efficiencies, under conditions of dilution ratio 7 and inlet NO concentration 1000 ppm. The results are summarized in Fig. 9.

Fig. 9 Effect of initial pH value of scrubbing solution on NO and NO_x removal efficiencies. Experimental conditions: dilution ratio (seawater/electrolyte) 7, inlet NO concentration 1000 ppm.

When the initial pH value of scrubbing solution increased from 3 to 5, the NO removal efficiency maintained 100%, while the NO_x removal efficiency decreased gradually. However, when the initial pH value of scrubbing solution was in the range of 6–8, the NO and NO_x removal efficiencies at pH value of 7 reached the peak value of 80.7% and 60.9%, respectively. Then, the NO and NO_x removal efficiencies decreased sharply with the pH value further increasing to 9. The effect of initial pH value on NO and NO_x removal efficiencies by the electro-generated chlorine solution can be explained by the impacts of pH on the distribution of active chlorine species. With pH changing, there may be four kinds of oxidizing agents in the scrubber, which can be represented by:

The reactions of dissolution and hydrolysis of chlorine in acid seawater solution are shown as follow:

Cl₂(aq) + H₂O ↔ HOCl + H⁺ + Cl⁻ (10)

Cl₂(aq) + Cl⁻ ↔ Cl₃⁻ (11)

p_Cl2(aq) = K_H[Cl₂(aq)] (12)

[ACC]_T = [Cl₂(aq)] + [Cl₃⁻] + [HOCl] + [OCl⁻] (13)

where K_H is the Henry's constant (16.1 atm L mol⁻¹ (25 °C)); [ACC]_T is the total concentration of chlorine in solution (mg L⁻¹ [Cl₂]).

Considering the ionic strength effect of scrubbing solution (ionic strength I ≈ 0.49 M), the corrected equilibrium constants for dissociation of hypochlorous acid and hydrolysis of chlorine reactions are represented by:³²

Then the relative molar percentage of Cl₂(aq), HOCl and OCl⁻, as well as the partial pressure of gaseous chlorine Cl₂(g) can be expressed in eqn (14)–(18).

As the seawater has a high concentration of Cl⁻ ion, we proposed the hypothesis that the Cl⁻ ion concentration unchanged during the experiments. Therefore, the pH-dependant active chlorine species and gaseous chlorine were calculated and plotted in Fig. 10.

Fig. 10 Distribution of active chlorine species as a function of pH in seawater. Experimental condition: dilution ratio (seawater/electrolyte) 7.

According to the literature,^{12,13,18,33,34} reaction pathways of NO absorption by electro-generated chlorine solution were proposed based on the pH-dependant active chlorine species and gaseous chlorine (Fig. 11).

Fig. 11 Proposed reaction pathways of NO and active chlorine reaction system.

When the pH value was below 5, NO_x gas was mainly removed by HOCl, Cl₂(aq) and Cl₂(g) in both liquid and gas phase. HOCl and Cl₂(aq) were strong oxidants, which contributed to a high removal efficiency of NO gas (reactions (19)–(21), (24) and (27)). Moreover, some of the aqueous molecular chlorine escaped into the gas phase in the scrubber, and then reacted directly with NO gas (reactions (28) and (29)). However, when the pH value increased from 3 to 5, the concentrations of Cl₂(aq) and Cl₂(g) decreased, while the concentration of HOCl increased. In this case, the concentration of NO₂(aq) produced by HOCl (reaction (21)) increased, which caused the reduction of NO_x removal efficiency.

However, when the pH value was greater than or equal to 6, NO_x gas was mainly absorbed by HOCl and OCl⁻ ion. The concentration of HOCl decreased, whereas the concentration of OCl⁻ increased when the pH value changed from 6 to 9. As the standard redox potential of HOCl was lower than that of OCl⁻, NO removal efficiency was expected to decrease with the increase of pH value of scrubbing solution. However, it was interesting to note that both NO and NO_x removal efficiencies reached a peak value at pH value of 7 of scrubbing solution. This might be attributed to two reasons. On the one hand, the NO₂ absorption efficiency by HOCl/OCl⁻ was not as efficient as the NO oxidation efficiency by HOCl/OCl⁻.²⁹ Thus, although HOCl was strong oxidants which converted NO(aq) into NO₂(aq), the high concentration of NO₂(aq) might inhibit NO absorption by electro-generated chlorine solution and escape into the gas phase. Nevertheless, it was worth noting that an increase of OH⁻ ion concentration was believed to promote the production of NO₂⁻ and NO₃⁻ by NO₂(aq) absorption reactions (22)–(23) and (25)–(26). Then, this NO₂⁻ further converted to NO₃⁻ by an excessive amount of HOCl in solution through reactions (24), which in turn accelerated NO₂(aq) absorption. Therefore, when the pH value changed from 6 to 7, NO and NO_x removal efficiencies increased from 76% and 55.3% to 80.8% and 60.9%, respectively. On the other hand, according to results of hypochlorous acid decomposition reported by Luke C. Adam et al.,³⁵ the HOCl had a maximum decomposition rate at about 7. The intermediates, such as ClO₂⁻ and Cl₂O·H₂O, might react with NO(aq), and then enhance NO and NO_x removal efficiencies at pH value of 7.

When the pH value reached 9, the molar percentage of OCl⁻ in electro-generated chlorine solution was over 90%. NO_x gas was mainly absorbed by OCl⁻ through reactions (30) and (31). As the NO absorption rate by OCl⁻ was lower than that by HOCl, the NO and NO_x removal efficiencies decreased as pH value increased. The NO₂(aq) generated by oxidation of NO gas was completely absorbed in this alkaline solution.

NO(g) ↔ NO(aq) (19)

NO(aq) + Cl₂(aq) + H₂O ↔ NO₂(aq) + 2H⁺ + 2Cl⁻ (20)

NO(aq) + HOCl ↔ NO₂(aq) + H⁺ + Cl⁻ (21)

3NO₂(aq) + H₂O ↔ 2H⁺ + 2NO₃⁻ + NO(aq) (22)

2NO₂(aq) + H₂O ↔ HNO₂ + H⁺ + NO₃⁻ (23)

NO₂⁻ + HOCl → NO₃⁻ + H⁺ + Cl⁻ (24)

HNO₂ ↔ H⁺ + NO₂⁻ (25)

3HNO₂ ↔ H⁺ + NO₃⁻ + 2NO(aq) + H₂O (26)

NO₂(aq) ↔ NO₂(g) (27)

2NO(g) + Cl₂(g) → 2NOCl(g) (28)

NOCl(g) + H₂O → H⁺ + Cl⁻ + HNO₂ (29)

NO(aq) + OCl⁻ ↔ NO₂(aq) + Cl⁻ (30)

NO₂⁻ + OCl⁻ → NO₃⁻ + Cl⁻ (31)

4. Conclusions

In this study, a method of NO_x removal by seawater electrolysis with less energy consumption was proposed. The following conclusions can be drawn:

(1) The active chlorine concentration increased linearly with the increase of current density and electrolysis time. The active chlorine formation rate kept a linear model with current density, which contributed to quantitative prediction and automatic control of active chlorine production onboard. The energy consumption decreased from 83 to 6.1 kW h per (kg [Cl₂]) with the salinity varying from 3.2 to 37.8 ppt after 5 min of electrolysis. Electrolysis using concentrated seawater from desalination plants operated on a ship was suggested to be a cost-effective solution.

(2) The effect of pH value had a substantial impact on NO_x removal performance by electro-generated chlorine, followed by the effect of dilution ratio (seawater/electrolyte) and inlet NO concentration. Results showed that the NO_x removal efficiency at pH value of 7 decreased from 98.9% to 20.2% with the dilution ratio varying from 2.4 to 96. In contrast, the NO_x removal efficiency without pH control reduced from 62.2% to 6.57% with the dilution ratio varying from 1 to 32. These pH-dependent results suggested that HOCl was the more active than OCl⁻ in terms of NO absorption. Compared with the inlet SO₂ concentration, the NO absorption process was more sensitive to inlet NO concentration. The reaction mechanism and pathways of NO_x removal by electro-generated chlorine were speculated by discussing the pH-dependant distribution of active chlorine species.

Further experiments are suggested to calculate the kinetics and model the absorption process, so that more engineering valuable information can be obtained to design and optimize the exhaust gas cleaning system for NO_x emission control onboard.

Acknowledgements

This paper was supported by Science and Technology Plan Project of China's Ministry of Transport (2015328225150) and the Fundamental Research Funds for the Central Universities (3132016018), and the Scientific Research Fund of Liaoning Provincial Education Department of China (No. L2014198).

Notes and references

1. J. Tournadre, Geophys. Res. Lett., 2014, 41, 7924–7932.
2. Z. Wan, M. Zhu, S. Chen and D. Sperling, Nature, 2016, 530, 275–277.
3. T. W. P. Smith, J. P. Jalkanen, B. A. Anderson, J. J. Corbett, J. Faber, S. Hanayama, E. O'Keeffe, S. Parker, L. Johansson, L. Aldous, C. Raucci, M. Traut, S. Ettinger, D. Nelissen, D. S. Lee, S. Ng, A. Agrawal, J. J. Winebrake, M. Hoen, S. Chesworth and A. Pandey, Third IMO GHG Study 2014, International Maritime Organization, London, UK, 2014.
4. H. Liu, M. Fu, X. Jin, Y. Shang, D. Shindell, G. Faluvegi, C. Shindell and K. He, Nat. Clim. Change, 2016, 6, 1037–1041.
5. IMO, MARPOL Annex VI and NTC 2008 with Guidelines for Implementation, International Maritime Organization, London, UK, 2013 edn, 2013.
6. D. J. Blatcher and I. Eames, Mar. Pollut. Bull., 2013, 74, 10–18.
7. Y. Sun, E. Zwolińska and A. G. Chmielewski, Crit. Rev. Environ. Sci. Technol., 2016, 46, 119–142.
8. A. Azzara, D. Rutherford and H. Wang, Feasibility of IMO Annex VI Tier III implementation using Selective Catalytic Reduction, International Council on Clean Transportation, 2014.
9. S. Cimino, L. Lisi and M. Tortorelli, Chem. Eng. J., 2016, 283, 223–230.
10. R.-T. Guo, J.-K. Hao, W.-G. Pan and Y.-L. Yu, Sep. Sci. Technol., 2015, 50, 310–321.
11. E. Ghibaudi, J. R. Barker and S. W. Benson, Int. J. Chem. Kinet., 1979, 11, 843–851.
12. L. Chen, C. H. Hsu and C. L. Yang, Environ. Prog., 2005, 24, 279–288.
13. R. Guo, W. Pan, X. Zhang, H. Xu, Q. Jin, C. Ding and S. Guo, Sep. Sci. Technol., 2013, 48, 2871–2875.
14. M. K. Mondal and V. R. Chelluboyana, Chem. Eng. J., 2013, 217, 48–53.
15. C. V. Raghunath and M. K. Mondal, Asia-Pac. J. Chem. Eng., 2016, 11, 88–97.
16. S. An and O. Nishida, in 32^nd International Conference On Environmental Systems, SAE Technical Paper, San Antonio, Texas, 2002, vol. 1, pp. 2295–2303.
17. S. An and O. Nishida, JSME Int. J., Ser. B, 2003, 46, 206–213.
18. T.-W. Kim, S.-J. Choi, J.-H. Kim and J.-Y. Song, Korean Chem. Eng. Res., 2012, 50, 825–829.
19. Z. Han, S. Yang, D. Zheng, X. Pan and Z. Yan, SpringerPlus, 2016, 5, 1–12.
20. B. S. Oh, S. G. Oh, Y. Y. Hwang, H.-W. Yu, J.-W. Kang and I. S. Kim, Sci. Total Environ., 2010, 408, 5958–5965.
21. H. K. Abdel-Aal, S. M. Sultan and I. A. Hussein, Int. J. Hydrogen Energy, 1993, 18, 545–551.
22. H. D. Al-Hamaiedeh, Desalin. Water Treat., 2013, 51, 3521–3526.
23. H. Nath, X. Wang, R. Torrens and A. Langdon, J. Appl. Electrochem., 2010, 41, 389–395.
24. X. Y. Li, F. Ding, P. S. Y. Lo and S. H. P. Sin, J. Environ. Eng., 2002, 128, 697–704.
25. S. Trasatti, Electrochim. Acta, 2000, 45, 2377–2385.
26. Ø. Ulleberg, Int. J. Hydrogen Energy, 2003, 28, 21–33.
27. N. Voutchkov, Desalination, 2011, 273, 205–219.
28. J. C. Morris, J. Phys. Chem., 1966, 70, 3798–3805.
29. H. Kobayashi, N. Takezawa and T. Niki, Environ. Sci. Technol., 1977, 11, 190–192.
30. C. N. James, R. C. Copeland and D. A. Lytle, in AWWA Water Quality Technology Conference, San Antonio, TX, 2004.
31. Y. Liu, J. Zhang, F. Xie, Q. Wang, J. Pan and Y. Yin, Chem. Eng. Res. Des., 2016, 111, 196–203.
32. Black and Veatch, White's Handbook of Chlorination and Alternative Disinfectants, Wiley, Hoboken, New Jersey, 5th edn, 2010.
33. P. Z. Wang, P. K. Cen, P. J. Zhou and P. J. Fan, in Simultaneous Multi-Pollutants Removal in Flue Gas by Ozone, Springer, Berlin Heidelberg, 2014, pp. 71–93.
34. Y. Zhou, C. Li, C. Fan, M. Fu, L. Tao, M. Yu and M. Zhang, Environ. Prog. Sustainable Energy, 2015, 34, 1586–1595.
35. L. C. Adam, I. Fabian, K. Suzuki and G. Gordon, Inorg. Chem., 1992, 31, 3534–3541.

---