Electrochemically-assisted ammonia recovery from wastewater using a floating electrode

The electrolytic decomposition of ammonia has been undertaken in solutions of varying chloride concentration using Ti/RuO₂-Pt electrodes as anode and Ti as cathode (Ding et al. 2010). Current densities in the range of 5–15 mA.cm⁻² were applied and ammonia removal increased with increasing current density and chloride ion addition. An optimum ammonia removal was achieved using 0.3 g l⁻¹ NaCl and a current density of 10 mA cm⁻². Under these conditions, ammonia was found to oxidise to nitrite and then further to nitrate. Some denitrification (NO₃⁻ conversion to N₂) was thought to occur (at the cathode). Li et al. used modelling to predict that optimal ammonia removal could be achieved using 0.31 g l⁻¹ NaCl, 42.75 mA cm⁻² for 101 minutes, at which point the ammonia and nitrate concentrations would be reduced to 0 and 1.1 mg l⁻¹, respectively (Li et al. 2011).

An electrochemical cell was constructed from 110 mm diameter PVC tubing, which was cut to a length of 135 mm and sealed with a circular base and air-tight screw cap (further detail is available in the Supplementary Information, Figure S1, available with the online version of this paper). Electrodes were circular sections of 0.1 cm thick ruthenium/iridium oxide coated titanium mesh (CC87, Electrode Coating Technologies, Australia) with a radius of 5.5 cm. The active electrode surface area for each electrode was estimated, using grey-scale thresholding of an optical image, to be 111.0 cm² (Schneider et al. 2012). During experiments, the working electrode (WE) was polarised cathodically while the counter electrode (CE) acted as the anode. A silver/silver chloride reference electrode was mounted midway between the WE and CE electrodes.

The cell was filled with 570 ml of wastewater solution, the air outlet was connected to a one litre glass bubbling cylinder filled with 570 ml of dilute H₂SO₄ solution, and the air inlet was connected to a mass flow controller (Omega, USA) supplying ambient air at 1 or 20 L min⁻¹. A constant current of 0, 1, 2 or 5 mA cm⁻² was applied between the electrodes using a Princeton Applied Research PAR273 potentiostat. 0.2 ml liquid samples were taken from the cell via a syringe port and from the H₂SO₄ solution at regular intervals.

An artificial inorganic wastewater solution based on OECD suggested guidelines was prepared containing 50 mM NH₄Cl, 1 mM CaCl₂·2H₂O, 1 mM MgSO₃·7H₂0, 15 mM KH₂PO₄, 5 mM Na₃PO₄·12H₂O and 2 mM NaCl (OECD 2001). It had a total nitrogen content of 0.05 M, conductivity of 7,000 mS m⁻¹ and pH of 7.0. Filtered anaerobic centrate from Suomenoja WWTP, Espoo, Finland was sampled and used within 24 h of sampling. The centrate had a typical concentration of 22 mg PO₄³-P l⁻¹, 49 mg Cl l⁻¹, 560 mg NH₄⁺-N l⁻¹, 0.2 mg NO₃⁻-N l⁻¹ and 0.2 mg NO₂⁻-N l⁻¹, conductivity of 630 mS m⁻¹ and pH of 7.4.

Ammonium, nitrate, nitrite, phosphate and chloride were measured using a Hach-Lange spectrophotometer with Hach-Lange test kits LCK303, LCK341, LCK339, LCK349 and LCK311, respectively.

Cathodic potentials were at maximum −4 V/Ag/AgCl, which suggests that oxidation of both chloride and water should occur. A faint smell of chlorine emanated from the receiving solution, accompanied by a slow reduction in chloride concentration that occurred with time (i.e. 53 mM to 46 mM over 8 hrs). The presence of nitrate and nitrite in the solution was determined to be 6 × 10⁻² mM and 4 × 10⁻¹ mM after 8 hrs, suggesting that low amounts of other water soluble oxidised forms of nitrogen are produced from (electro)chemical reactions (further data available in Supplementary Information, Figures S2 and S5, available with the online version of this paper).

These four processes are optimised where the cathode is floated on the surface of the wastewater. Ammonia removal was measured to be 215 mg l⁻¹ hr⁻¹ with a floating cathode (whereby the meniscus of the water met with the top of the electrode) and was shown to decrease to 68 mg l⁻¹ hr⁻¹ when immersed 0.6 mm in the solution (further detail is provided in Supplementary Information, Figure S3, available online). At the same time, the relative amount of ammonia recovered significantly decreased.

The influence of air flow rate on removal/recovery efficiency was also explored. Using identical conditions to those used to derive the data for Figure 2 but with an air flow rate of 5 L min⁻¹, no difference in total ammonia removal was observed (within error) as can be seen in Supplementary Information, Figure S4 (available online); however, the total amount of ammonia recovered at the reduced air flow rate was <5% of ammonia removed. In comparison, the ammonia recovery percentage at an air flow rate of 20 l min⁻¹ was 54 ± 8%. Further work is warranted to better understand the importance and control of air flow.

In order to assess the viability and transferability of the novel electrochemical approach to real wastewater, experiments were performed using filtered anaerobic centrate obtained from the Suomenojan wastewater treatment plant. An initial experiment was undertaken using a 5 mA cm⁻² electrochemical current and an airflow rate of 20 l min⁻¹. Under these conditions, there was considerable foam generation at the wastewater-air interface. The ammonia concentration decreased by approximately 110 mg l⁻¹ hr⁻¹ (6.5 mM hr⁻¹) from a starting concentration of 560 mg l⁻¹. This compares to removal efficiencies of 72 to 216 mg l⁻¹ hr⁻¹ (4 to 12 mM) under identical conditions from the synthetic wastewater composition, originally containing 900 mg l⁻¹. The varied removal rate from the latter represents the extremes of the electrode being slightly submerged at 0.6 mm and floating, respectively.

During treatment of the wastewater centrate, the levels of other ions were found to vary. As expected, the chloride ion concentration decreased upon the application of the electrochemical current. The chloride concentration in the centrate was considerably lower than that of the synthetic formulation (i.e. 1.4 mM versus 0.051 M). Nitrite levels were low both before and after treatment, maintaining levels below 0.002 mM. Nitrate levels were initially 0.1 mM and increased to 0.2 mM during the 3.5 hour experiment (see Supplementary Information, Figure S5, available online).

Ammonia oxidation can arise from either direct oxidation or through chemical oxidation. A faint smell of chlorine was noticed when using synthetic wastewater, suggesting the formation of Cl₂. Monitoring of chloride ion concentrations showed a slow loss from solution, confirming Cl₂ emission from the electrochemical cell. While the formation of Cl₂ is favoured at low pH (i.e. pH < 3.3) and would certainly exist in high concentrations around the anodic electrode, the major oxidant in the wastewater bulk is likely to be HClO (Equation (4)).

The influence of chloride on the electrochemical oxidation was not explored, however; slightly higher ammonia removal rates were achieved during experiments on the synthetic wastewater, which had high chloride concentrations (0.051 M) in comparison to the anaerobic centrate with its low chloride (0.0014 M). This is likely to be associated with the mechanisms described by Equations (3)–(6), leading to higher removal efficiencies at the anode.

Analysis of nitrite and nitrate during several experiments demonstrated that very low levels were obtained. Low rates of water oxidation, and also high alkalinity, act to minimise the generation of oxidised nitrogen compounds, whose formation would be detrimental for practical application. These observations match literature reports that suggest N₂ to be the main product from electrooxidation (Pressley et al. 1972).

The novel approach to ammonia removal and recovery at a floating cathodic electrode is demonstrated in Figure 3 to operate at higher efficiency than electrooxidation processes, showing recoveries of over 55%. The efficiency of electrochemical oxidation is typically evaluated against the three electron conversion of ammonia to N₂ in Equation (1) (Kim et al. 2006). The relationship between electrical energy and ammonia removal at the cathode is, however, complicated by the indirect mechanism of: (i) creation of hydroxyl ions, (ii) conversion of ammonium to ammonia, (iii) air-stripping of ammonia. The fundamental exchange of electrons to create alkalinity is for each electron to create one hydroxyl, which in turn may lead to one ammonia molecule being released into the air stream. This theoretically suggests that ammonia removal at the cathode may be three times more efficient than electrooxidation, and the combined process should require just two electrons to remove one ammonia molecule.

For the majority of experiments, the reduction in ammonia occurred linearly with time, suggesting concentration-independent removal. Several authors (Vanlangendonck et al. 2005; Li & Liu 2009) have demonstrated that the electrooxidation of ammonia from solution occurs linearly with time. Vanlangendonck et al. (2005) suggest this is due to the constant formation of HOCl and its subsequent ability to oxidise ammonia through Equations (3)–(6), which then regenerates chloride ions to maintain the constant reaction rate. It is also proposed that the applied electric field acts to transport positively-charged ammonium ions toward the negatively-charged cathode. This enables a consistently high concentration of ammonia to accumulate at the wastewater-air interface, allowing removal to occur.

The overall efficiency of electrical energy can be estimated by taking into account the theoretical requirement of two electrons to remove one ammonia molecule from solution. In the present work, 3.2 electrons were required to remove one ammonia molecule, which equates to a theoretical electrical efficiency of 63% and an energy cost of 29.3 kWh.Kg⁻¹ N to remove nitrogen. This efficiency compares well against other reported electrochemical treatments for wastewater (Kim et al. 2006; Desloover et al. 2015). The total energy cost exceeds that reported by Ippersiel et al. (2012) for combined electrodialysis and air-stripping (18 kWh.kg⁻¹ N removed), however, additional chemical costs to overcome low recovery efficiencies were not reported and accounted for by Ippersiel et al. (2012). Desloover et al. reported the ammonia concentration to be achieved using 13–18 kWh.kg⁻¹ N, with air stripping requiring a further 9 kWh.kg⁻¹ N to enable nitrogen recovery (Desloover et al. 2012, 2015).

When compared against reports of the relative efficiency of biological processes to remove N, it is apparent that electrochemical approaches are uncompetitive against biological deammonification processes, which are reported to remove ammonia with as little energy as 1.2 kWh.kg⁻¹ N (Wett et al. 2010). With regards to the value of ammonia through recovery, the Haber-Bosch process is reported as being an energy-intensive means to produce ammonia, requiring some 8 kWh.kg⁻¹ (De Beer 2000). The relatively low energy cost of producing bulk ammonia compared to the costs to remove ammonia using electrochemical approaches suggest that a step change in technology is required to warrant its recovery.

The practical utilisation of the EAAR process reported here will be dependent upon maximising its advantages while minimising disadvantages. The key advantages and disadvantages of electrooxidation processes have been documented by Anglada et al. (2009). Advantages include: robustness, versatility, low temperature and pressure requirements, and amenability to automation. Disadvantages are predominantly associated with the high cost of electrical energy consumption. The EAAR approach maintains the advantages listed above, and furthermore it demonstrates significant improvement in the efficiency of ammonia removal from wastewater over conventional electrooxidation processes, requiring 2:1 electron:ammonia rather than 3:1). The present approach also achieves ammonia recovery without the use of expensive ion-exchange membranes and/or chemical addition to achieve a high solution pH.

Additional disadvantages of EAAR are largely due to the placement of the cathodic electrode at the wastewater-air interface. The removal of ammonia from the bulk only occurs at the surface in a two-dimensional arrangement, which limits the volume of wastewater that can be treated by a set electrode area. This effect may however be less of a disadvantage due to the movement of ammonium ions toward the surface under the applied electric field. Further work is required to develop workable configurations for the process. An additional complication is the ability to float the cathodic electrode on the surface of the wastewater, which may require the development of novel low-density electrodes and/or engineered means to maintain optimal positioning of the electrode. Further advancement of the fundamental concepts of EAAR would benefit from greater knowledge of the influence of the cathodic electrode area and porosity, conductivity, chloride ion concentration and wastewater alkalinity. Finally, the influence of the air-flow rate required for optimal ammonia recovery requires further investigation, as does the cost associated with air flow.

Electrochemical methods for wastewater treatment are likely to be confined to decentralised treatment operations, where small scale treatment can be automated in a robust manner. One such example may be the localised treatment of urine or blackwater. In this case, the greater energy costs for electrochemical treatment may be justified by a saving in capital, maintenance and labour costs.

A novel method for the removal and recovery of ammonia from wastewater has been demonstrated through the floating of a cathodically-polarised porous electrode on the surface of wastewater.

The approach offers improved removal efficiency in comparison to electrooxidation, and furthermore, the combination of floating electrode and anodic electrooxidation mechanisms offer possible synergies in enhancing the overall efficiencies of electrochemical treatment of wastewater.

In comparison to existing electrochemical treatments, it also does not require the use of ion-exchange membranes in order to recover ammonia. The method utilises local pH increases, as opposed to the generation of bulk alkalinity.

Linear reductions of NH₄⁺ in wastewater were observed irrespective of concentration, suggesting the electric field creates a continuous flux of NH₄⁺ to the cathode-wastewater-air interface, where NH₃ can pass into the vapour phase.

Further development is required to better understand air flow requirements, electrode design and wastewater composition effects on efficiency.

The authors would like to acknowledge the CSIRO Julius Career Award scheme and Mona Arnold, VTT, for enabling this work to be undertaken.