Synthesis of Ammonia Directly from Wet Air Using Sm 0.6 Ba 0.4 Fe 0.8 Cu 0.2 O 3àd as the Catalyst

Ammonia was directly synthesised from wet air at 400 C at atmospheric pressure. A new perovskite Sm 0.6 Ba 0.4 Fe 0.8 Cu 0.2 O 3Àd was used as the electrocatalyst for electrochemical synthesis of ammonia. Ammonia formation rates of 9.19 Â 10 À7 mol s À1 m À2 and 1.53 Â 10 À6 mol s À1 m À2 were obtained at 400 C when wet air and wet N 2 were introduced into a simple single chamber reactor, respectively. The perovskite catalyst is low cost compared to the previously reported Ru/MgO and Pt/C catalysts. This experiment indicates that ammonia can be directly synthesised from wet air, a very promising simple technology for sustainable synthesis of ammonia in the future.


Introduction
Nitrogen fertiliser has supported approximately 27% of the world's population over the last century, equivalent to around 4 billion people born (or 42% of the estimated total births) since 1908. 1 Although over 78% of the atmosphere is composed of nitrogen, it is difficult to use N 2 directly from plants as plants can only use xed nitrogen.The breakthrough in nitrogen xation took place a century ago, the well-known Haber-Bosch process, operating at high temperatures ($500 C) and high pressures (150-300 bar). 2 Fritz Haber and Carl Bosch won the Nobel Prize in Chemistry in 1918 and 1931, respectively, for their contributions in industrial production of ammonia.This innovation became a key driver in the development of the industrialized society and is still securing the nutrition of billions of people today.In 2007, the Nobel Prize in Chemistry was awarded to Gerhard Ertl for his studies of chemical processes on solid surfaces.A major contribution to this prize was Gerhard Ertl's understanding of the reaction mechanism of H 2 and N 2 on solid ammonia synthesis catalysts. 3n the Haber-Bosch process, the required hydrogen is produced through the steam reforming of natural gases or coal of which consumed more than 1% of world power generation. 4In addition, CO 2 emitted from the ammonia industry is equivalent to 0.77% of the world total CO 2 emission.In order to reduce the current dependency on fossil fuels and to reduce carbon emissions associated with their use, it is essential to introduce new ammonia synthesis processes and break the link between fossil fuels and the ammonia industry. 5o reduce global CO 2 emission, renewable electricity generated from solar, wind, marine and other sources are becoming more and more important.Due to the intermittent nature of those renewable resources, energy storage, particularly renewable electricity storage has become a big challenge.One of the possible solutions for renewable electricity storage is to produce basic chemical feedstocks such as ammonia.7][8] Recently, it has been reported that ammonia can be directly synthesised from air and water at ambient conditions. 2 It has been widely reported that ammonia can be synthesised from N 2 and H 2 (or steam) at a temperature around 500 C. [9][10][11][12][13][14][15] The possible reason is that, the decomposition of ammonia at high temperature is kinetically slow thus ammonia can still be collected if it is removed from the reactor in time.Ammonia will be able to be synthesised from air and water at elevated temperatures if the oxidation of ammonia is also kinetically slow.
In our previous papers, we reported the electrochemical synthesis of ammonia based on electrochemical cells using a Ce 0.8 Sm 0.2 O 2Àd -(Li,Na,K) 2 CO 3 composite electrolyte. 12The eutectic point of (Li,Na,K) 2 CO 3 ternary molten salts is 396 C. 16,17 The low melting point of mixed (Li,Na,K) 2 CO 3 salts causes the composite to exhibit high ionic conductivity at relatively low temperatures which can minimise the operating temperature of the cell, reducing the thermal decomposition and oxidation of ammonia.It has been reported that the ionic conductivity of the Ce 0.8 Sm 0.2 O 2Àd -(Li,Na,K) 2 CO 3 composite reached 0.1 S cm À1 at a temperature around 400 C. 18 The total ionic conductivity includes those from Li + , Na + , K + , H + , HCO 3 À , CO 3 2À and O 2À ions.9][20] Either H + or O 2À ionic conduction can be used for electrochemical synthesis of ammonia. is a very good ammonia synthesis catalyst and an ammonia formation rate of 1.13 Â 10 À8 mol cm À2 s À1 was observed at 80 C when N 2 and H 2 were used as the reactants. 23This inspired us to use doped SmFeO 3 catalysts for the electrochemical synthesis of ammonia.On the other hand, BaO is a typical promoter of ammonia synthesis catalysts. 24,25However, BaO itself is unstable in wet air because of the reaction between BaO and H 2 O to form Ba(OH) 2 .Therefore, in this paper, we introduce some Ba at the A-site of perovskite oxide SmFeO 3 .A new perovskite, Sm 0.6 Ba 0.4 Fe 0.8 Cu 0.2 O 3Àd (SBFCu), was synthesised as an electro-catalyst for the synthesis of ammonia from air/N 2 and water.Sm 0.6 Ba 0.4 Fe 0.8 Cu 0.2 O 3Àd was used as both the cathode and anode and a single chamber reactor was used for the synthesis.This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

Materials synthesis
Calculated amounts of Gd 2 O 3 were dissolved in hot dilute nitric acid to form a gadolinium nitrate aqueous solution.Then an appropriate amount of Ce(N-O 3 ) 3 $6H 2 O was added to the solution.A 1 M (NH 4 ) 2 CO 3 solution was added dropwise with vigorous stirring until the pH value reached 7-8, aer that vigorous stirring was continued for another 30 min.The formed precipitate was then collected by vacuum ltration and washed several times with deionised water then dried at 100 C overnight.The obtained powder was calcined in air at 650 C for 2 hours with a heating/cooling rate of 2 C min À1 .Finally, ultrane powder of GDC was obtained.Ce 0.8 Gd 0.2 O 2Àd and ternary carbonate mixture (Li,Na,K) 2 CO 3 (32.1 wt% Li 2 CO 3 ; 34.5 wt% K 2 CO 3 ; 33.4 wt% Na 2 CO 3 ) powders were mixed together with a weight ratio of oxide to carbonate of 7 : 3. The mixture was mixed and ground with the use of 25 ml isopropanol.Then the mixture was put in a ball miller (Planetary Mono Mill Pulverisette 6) for mixing with a speed of 400 rpm for 4 hours.The weight ratio of balls to powders was roughly 1 : 1.Then the mixture was heated by a hot plate magnetically stirred at 150 C to release and remove the isopropanol.The powder was heated at 600 C for one hour then quenched in air to room temperature.The as-prepared oxide-carbonate composite electrolyte will be used for ammonia synthesis.
Samarium(III) oxide was dissolved in diluted nitric acid to form samarium nitrate under heating and stirring.Calculated amounts of nitrates Ba(NO 3 ) 2 , Fe(NO 3 ) 3 $9H 2 O and Cu (NO 3 ) 3 $2.5H 2 O were dissolved in deionised water and were added to the above prepared solution.Appropriate amounts of citric acid and EDTA (ethylenediaminetetraacetic acid) were then added as complexing agents with a molar ratio of citric acid : EDTA : metal cations of 1.5 : 1 : 1. Dilute aqueous ammonia solution was then added to the mixed solution to adjust the pH value to around 6 and a dark green clear solution was obtained.The mixed solution was evaporated on a hot-plate at 200-250 C under stirring which gradually changed into a dark sticky gel before complete drying.By further heating the gel converted to a black colour.The brown ash was grounded and subsequently calcined in air at 900 C for 2 hours with heating/cooling rates of 5 C min À1 to obtain a single phase of Sm 0.6 Ba 0.4 Fe 0.8 Cu 0.2 O 3Àd .

Materials characterisation
X-ray data were collected on a PANanalytical X' Pert Pro in the Bragg-Brentano reection geometry with a Ni-ltered Cu Ka source (1.5405 Å), tted with a X'Celerator detector and an Empyrean CuLFF XRD tube.Absolute scans in the 2q range of 5-100 with step sizes of 0.0167 were used during data collection.The surface area of Sm 0.6 Ba 0.4 Fe 0.8 Cu 0.2 O 3Àd powder was measured by a N 2 adsorption method using Micromeritics ASAP 2420.
TG/DSC analyses were performed using a Stanton Redcro STA/TGH series STA 1500 operating through a Rheometric Scientic system interface controlled by the soware RSI Orchestrator in owing air or N 2 at a ow rate of 50 ml min À1 .
Total conductivity measurements were carried out by a pseudo-four-terminal method using a computer-controlled Solartron Analytical® SI 1470A electrochemical interface by applying a constant current.The Sm 0.6 Ba 0.4 Fe 0.8 Cu 0.2 O 3Àd powder red at 900 C was pressed into pellets with diameters of 13 mm and thicknesses of around 2 mm which were then red at 1100 C for 5 hours.A silver coated pellet was tted into the measuring apparatus and the conductivity measurement was carried out in air between room temperature and 750 C.

Fabrication of the single cell and ammonia synthesis measurements
The cell was placed in a self-designed single chamber reactor as described elsewhere. 22The electrolytic cell for ammonia synthesis was a tri-layer single cell which was fabricated by a cost-effective one-step dry-pressing and co-ring process.The anode and cathode material was 0.8 g Sm 0.6 Ba 0.4 Fe 0.8 Cu 0.2 O 3Àd plus 0.2 g composite electrolyte to form composite electrode materials.The composite anode, composite electrolyte and composite cathode (SBFCu/CGO-(Li,Na,K) 2 CO 3 / SBFCu) were fed into the die, layer by layer, with the aid of a sieve to ensure uniform powder distribution, and then unixally pressed at a pressure of 300 MPa by cold pressing into 13 mm pellets.The pellets were sintered at 600 C for 2 h.The thickness of the anode, electrolyte and cathode was approximately 0.6, 1.0 and 0.6 mm, respectively.The catalyst surface area of the cathode and anode was 1.281 cm 2 .Silver paste was painted on each electrode surface of the cell as a current collector.Ag wires were used as output terminals for both electrodes.It has been reported that silver itself has negligible catalytic effects on ammonia synthesis. 9ompressed air or N 2 was passed through room temperature water before inputting into the single chamber reactor as described in previous reports. 21,22he ow rate of air (or N 2 ) was 50 ml min À1 .The steam concentration in air (or N 2 ) was 3 mol%.A DC voltage was applied to the cell using a Solartron 1287A electrochemical interface.The synthesised ammonia was absorbed by 25 ml of diluted sulphuric acid (0.001 M) for 30 min as described before. 2,12The concentration of NH 4 + in the absorbed solution was analysed using the Orion Application solution for low adjusting ISA.The produced ammonia was detected using an ammonia meter (ISE Thermo Scientic Orion Star A214) and the rate of ammonia formation was calculated using the following equation.
where: [NH 4 + ] ¼ is the measured NH 4 + ion concentration, V ¼ is the volume of the dilute H 2 SO 4 for ammonia collection, t ¼ is the adsorption time and A ¼ is the effective area of the cell.

Results and discussion
Thermal analysis of the brown ash was carried out in air (Fig. 1A).It was found that a weight loss of $52% occurred at a temperature between 300 and 430 C.This is due to the decomposition of the ash and loss of organic components.The sample weight became stable at a temperature above 800 C indicating the minimum formation temperature for Sm 0.6 Ba 0.4 Fe 0.8 Cu 0.2 O 3Àd is 800 C. To guarantee the formation of a single phase, we choose 900 C as the ring temperature for Sm 0.6 Ba 0.4 Fe 0.8 Cu 0.2 O 3Àd .The ash was red at 900 C for 2 hours to form a perovskite phase.XRD analysis of the powder aer ring at 900 C indicates it is a single phase (Fig. 2).It exhibits an cubic structure with a space group of Pm 3m(221); a ¼ 3.9148(1) Å, V ¼ 60.00(1) Å3 (Table 1).In our previous report, perovskite oxide La 0.6 Sr 0.4 Fe 0.8 Cu 0.2 O 3Àd exhibits a cubic structure with a space group of Pm-3m(221). 12SEM pictures indicate that the prepared Sm 0.6 Ba 0.4 Fe 0.8 Cu 0.2 O 3Àd powders agglomerated together with a secondary particle size of $0.1 mm (Fig. 3).The surface area of the La 0.6 Sr 0.4 Fe 0.8 Cu 0.2 O 3Àd prepared at 900 C was 7.94(1) m 2 g À1 .
The electronic conductivity of Sm 0.6 Ba 0.4 Fe 0.8 Cu 0.2 O 3Àd is very important in order to be used as both the cathode and anode for the electrochemical cells for ammonia synthesis.The DC conductivity measurement indicated that the conductivity increased before 550 C, reaching a value of 2.29 S cm À1 and then started to decrease (Fig. 4).This is probably due to the semi-conductor to metal transition which is a common phenomenon in perovskite oxides. 26The conductivity of Sm 0.6 Ba 0.4 Fe 0.8 Cu 0.2 O 3Àd was 1.88 S cm À1 at 400 C. Regarding the operating temperature of the cell, the lowest working temperature of the CGO- For ammonia synthesis, wet air was fed into the single chamber reactor as described elsewhere. 22The gas composition was approximately 3 mol% H 2 O, 97 mol% air.The effect of the applied potential on ammonia formation rate was investigated at 400 C and the order for applied voltage was from low to high, between 1.2 and 1.7 V with and interval of 0.1 V. Fig. 5A shows the recorded current change against time for the electrolytic cell during the ammonia synthesis process at different potentials at 400 C over a period of 3600 seconds (was 3300 seconds due to an interruption when applying 1.6 V).As can be seen, the current density decreased at the beginning but tended to become stable.The ammonia  produced in the single chamber reactor was absorbed by 20 ml of dilute H 2 SO 4 (0.001 M) for 60 minutes.It is surprising that ammonia was successfully synthesised at atmospheric pressure from wet air ($3% H 2 O).As can be seen in Fig. 6A the ammonia formation rate increased signicantly with increasing the applied potential and reached a maximum value at 1.4 V (9.19 Â 10 À7 mol s À1 m À2 at 400 C), showing that 1.4 V was the optimum potential for ammonia synthesis in this study.In terms of catalyst weight, the ammonia formation rate was 1.48 Â 10 À10 mol s À1 g cat À1 .This is two orders of magnitude higher than the reported ammonia formation rates (around 1.0 Â 10 À12 mol s À1 g cat À1 at $600 C) when either a H + (SrCe 0.95 Yb 0.05 O 3Àd ) or O 2À (8 mol% yttria-stabilized zirconia) conductor was used as the electrolyte with an industrial Ru/MgO catalyst used at the cathode. 9It was 400 C in our experiment which is roughly 200 C lower than that reported by Skodra. 9Higher working temperatures may cause the thermal decomposition or oxidation of the produced ammonia, particularly at a temperature above 500 C. 2 The Sm 0.6 Ba 0.4 Fe 0.8 Cu 0.2 O 3Àd catalyst used in this study is much cheaper than the Ru/MgO catalyst reported by Skodra and Stoukides 9 and the Pt/C catalyst used for low temperature synthesis of ammonia from air and water. 2,4or comparison, ammonia was synthesised from wet nitrogen using the same cell.Fig. 5B shows the recorded current change against time for the electrolytic  cell during the ammonia synthesis process at different potentials at 400 C over a period of 60 minutes.As can be seen, the electrolytic cell was stable when the applied voltage was over 1.4 V.As expected, ammonia was also successfully synthesised at 400 C under atmospheric pressure from $3% H 2 O and N 2 (Fig. 6A).
Again the ammonia formation rate increased signicantly by increasing the applied potential and reached a maximum value at 1.4 V (1.53 Â 10 À6 mol s À1 m À2 at 400 C), showing that 1.4 V was the optimum potential for ammonia synthesis in this study.The same tendency was observed for both wet air and wet N 2 .This could be attributed to nitrogen chemisorption hindered by the high rate of electrochemically supplied H + , which in turn poisoned the catalyst (cathode surface). 6,27In terms of catalyst weight, the ammonia formation rate was 2.45 Â 10 À10 mol s À1 g cat À1 which is more than two orders of magnitude higher than the values reported by Skodra. 9When applying 1.7 V, the ammonia formation rates for wet air and wet N 2 are comparable (Fig. 6A) while slightly more ammonia was produced at lower voltages when wet N 2 was introduced into the single chamber reactor.The observed ammonia formation rates in this study were lower than the reported value when SmFe 0.7 Cu 0.1 Ni 0.2 O 3Àd was used as the cathode catalyst but the experiments were operated under different conditions. 23H 2 and N 2 were used as the reactant while we used wet N 2 /air as the precursor.The cell operating temperature was 400 C while it was near room temperature in a previous report. 23he Faraday efficiencies of the electrochemical synthesis process under different conditions are also calculated.As shown in Fig. 6B, when wet air was used, the highest Faraday efficiency was 0.74% when a voltage of 1.7 V was applied.In the case of wet N 2 , the highest Faraday efficiency was 1.19% at a voltage of 1.7 V.At lower applied voltages, the Faraday efficiencies were comparable for wet air and wet N 2 , indicating the possibility to use wet air instead of wet N 2 for direct synthesis of ammonia at temperature or below 400 C.Although the Faraday efficiencies in our experiments are lower than the high efficiency of 78% when H 2 and N 2 was used as the precursor, 13 steam instead of H 2 , and air instead of N 2 were used as the precursors in this study.It is believed that both the formation rate and Faraday efficiency can be improved when better electro-catalysts are identied.In conclusion, aer our previous reports on ammonia synthesis directly from air and water at ambient temperature up to 80 C, 2,4 ammonia was also synthesised directly from wet air at a higher temperature (400 C).Perovskite oxide, Sm 0.6 Ba 0.4 Fe 0.8 Cu 0.2 O 3Àd , was successfully synthesised and showed a conductivity of 1.88 S cm À1 at 400 C. Ammonia was successfully synthesised at 400 C under atmospheric pressure from both wet air and wet nitrogen using Sm 0.6 Ba 0.4 Fe 0.8 Cu 0.2 O 3Àd as the anode and cathode catalyst.Ammonia formation rates of 9.19 Â 10 À7 mol s À1 m À2 and 1.53 Â 10 À6 mol s À1 m À2 were obtained at 400 C when wet air and wet N 2 were introduced into a simple single chamber reactor, respectively.These values are more than two orders of magnitude higher than the reported ammonia formation rates when synthesised from N 2 and H 2 O at 600 C. 9 Our perovskite catalyst is also low cost compared to the Ru/MgO catalyst in a previous report 9 and Pt/C catalysts in our previous study on ammonia synthesis from air and water. 2,4Although the ammonia formation rates were not high enough for mass production, this is a very simple process using low-cost materials.Aer further investigation and optimisation, we believe ammonia can be synthesised directly from wet air or wet N 2 on a large scale to feed the growing world population in the future.

Fig. 1
Fig. 1 The STA analysis of the Sm 0.6 Ba 0.4 Fe 0.8 Cu 0.2 O 3Àd ash (A) and the Sm 0.6 Ba 0.4 Fe 0.8 Cu 0.2 O 3Àd powder in N 2 (B).

Fig. 5
Fig.5The current density across the cell in wet air (A) and wet N 2 (B) when applied with different voltages.

Fig. 6
Fig.6The ammonia formation rate (A) and Faraday efficiency (B) at 400 C in wet air and wet N 2 when applied with different voltages.

6 Therefore a similar composite electrolyte Ce 0.8 Gd 0.2 O 2Àd -(Li,Na,K) 2 CO 3 was used as the electrolyte in this study for ammonia synthesis. Recently, it has been reported that
ammonia was successfully synthesised directly from wet air using La 0.8 Cs 0.2 Fe 0.8 Ni 0.2 O 3Àd 21 or Pr 0.6 Ba 0.4 Fe 0.8 Cu 0.2 O 3Àd 22 as the catalyst.It has been reported that Cu and Ni co-doped SmFeO 3 , SmFe 0.7 Cu 0.1 Ni 0.2 O 3Àd