1 Introduction

Ammonia plays an important role in human life and industrial production because of its wide application in fiber, chemical fertilizer, pharmaceutical, energy storage intermediate and other fields [1]. Nowadays, available ammonia preparation methods can be classified into industrial Haber-Bosch, natural biological, plasma discharge, photocatalytic and electrochemical nitrogen fixation process. Although the traditional Haber-Bosch reduction process has dominated the industrial scale NH3 synthesis field, it usually suffers from low efficiency, high fossil energy consumption and severe operation conditions (ca. 700 K and ca. 100 atm) even with some optimized catalysts due to the stable triple bond of nitrogen molecules (945 kJ mol− 1), low polarization level and high ionization energy [2, 3]. Therefore, it remains a great challenge to look for effective methods to promote the industrial scale ammonia production via effectively adsorbing and activating nitrogen [4, 5]. Meanwhile, biological nitrogen fixation cannot meet human needs for NH3 development due to its low yield, uncontrollable reaction process and slow reaction rate [6]. Comparatively speaking, photocatalytic and electrochemical nitrogen fixation are two sustainable and green methods for ammonia synthesis due to their ambient working conditions and low energy consumption mode, and the electrochemical nitrogen fixation process shows a higher ammonia yield, easier handling conditions and wider application field [7].

The biggest challenge of electrochemical ammonia synthesis lies in the search and design of highly effective electrocatalysts to activate the ultra-highly stable three bonds of nitrogen molecules, promote N2 reduction reaction (NRR) kinetics and weaken hydrogen evolution reaction (HER) interference. As a suitable catalyst for NRR, it should satisfy four prerequisites: high ammonia yield rate, satisfactory faraday efficiency, excellent activity and good selectivity. In general, various design strategies for the reported catalysts can be classified into four categories, that is, defect engineering, structural manipulation, crystallographic tailoring, and interface regulation [6]. So far, a series of effective catalysts have been prepared to electrochemically reduce nitrogen to ammonia, which contain precious metals (e.g., Ru [8], Rh [9], Au [10] and Pt [11], etc.), non-precious metals compounds (e.g., Fe [12], foam nickel [13], MoS2 [14], VN [15], etc.) and metal-free nanomaterials (e.g. C3N4 [16], B-doped graphene [17], N-doped carbon [18], B4C [19], black phosphorus nanosheets [20], etc.). Among these effective catalysts, non-precious metal Fe-based nanomaterials presented a promising catalytical effect on the electrochemical nitrogen reduction, which exhibited an outstanding activity and feasibility for electrocatalytic NRR [21,22,23,24]. For example, inspired by the role of biological nitrogenase, FeMo nanoclusters/single atoms was designed and synthesized on porous nitrogen-doped carbon (FeMo/NC) [25], indicating a stable Faradaic efficiency (FE) (ca. 11.8%) at -0.25 V and NH3 yield rate (ca. 26.5 µgh−1 mg−1cat.) at -0.3 V in neutral electrolyte. Moreover, the appearance of the single-atom Fe on nitrogen-doped carbon can achieve a maximum Faradaic efficiency of 56.55% and special positive initial potential of 0.193 V for NH3 production [26]. The single Fe atoms supported on nitrogen-doped carbon can bring a FE of 8.4% at -0.2 V [27]. As reported [24], the optimized charge transfer between adjacent O and Fe atoms can both promote the electroreduction kinetics of N2 to NH3, and effectively reduce the binding energy between atomic Fe catalyst and electrochemical reaction intermediates of N2. Wang et al. found the porous Fe2O3 nanorods supported on carbon cloth provided a high NH3 yield of 6.78 µg h− 1 cm− 2 and Faradaic efficiency of 7.69% at -0.4 V [28]. The NH3 yield that catalyzed by the optimized Fe doped CuS QDs reached to 26.4 µg h−1 mg−1cat at -0.7 V, which quintuples that of the pristine CuS QDs [29]. As proved, nitrogen can be easily captured at ambient temperature by the Fe atoms that incorporated into the graphene layer, exhibiting a special six-proton and six-electron process [30]. More importantly, benefiting from the empty d orbitals of transition metals atoms, Fe-based catalysts can synergistically accept electrons from nitrogen gas molecules with appropriate electronic energy and symmetry [31]. At the same time, metal Fe atoms act as electron acceptors to donate electrons to the π* orbitals of nitrogen molecules, strengthening their adsorption and activating the triple bonds of nitrogen molecules [32]. Most of these Fe-based catalysts are porous carbon supported single-atom Fe catalyst, and their stability and practical application prospect are questionable. In this work, one novel Fe-based complex catalyst, Fe-tetrapyridophenazine (Fe-TPPHZ) was synthesized and utilized to electrochemically catalyze nitrogen. The TPPHZ can be used as one ideal protective agent for metal catalysts, which can restrain them from aggregating and deactivating. Furthermore, the nitrogen atoms of the TPPHZ help to boost the ambient nitrogen reduction reactions [33].

2 Experimental section

2.1 Chemicals and reagents

All chemicals were used as obtained without further purification. Tetrapyridophenazine (TPPHZ, 98 wt%) was purchased from Jinan Henghua Technology Co., Ltd. (China). N-N-dimethylformamide (DMF, 99.5%), dichloromethane, isopropanol, ethanol, ferrous sulfate heptahydrate, potassium hydroxide, sodium hydroxide, sulfuric acid, hydrochloric acid, sodium salicylate, potassium sodium tartrate, sodium hypochlorite, ammonium chloride, p-Dimethylaminobenzaldehyde, hydrazine hydrochloride and sodium nitroprusside (99.0 wt%) were bought from Sinopharm Chemical Reagent Co., Ltd. (China). Keqin black (KB) comes from Suzhou Yilongsheng Energy Technology Co., Ltd (China). Nafion solution (5 wt%) was provided by Nation DuPont Co., Ltd. (USA).

2.2 Synthesis of Fe-TPPHZ

Fe-TPPHZ was synthesized via a hydrothermal method [34]. Firstly, 0.104 g (0.25 mmol) TPPHZ was put into a 70 mL DMF solution and jointly ultrasonicated for 10 min to form a uniform solution. Secondly, 0.07 g (0.25 mmol) FeSO4·7H2O was added into the above mixture under vigorous stirring and then they were transferred to a 100 mL Teflon lined stainless steel autoclave together. After being hydrothermally heated at 160 ℃ for 12 h, the sample was handled by a series of centrifugation and washing procedures. In the rinse process, the unreacted TPPHZ was washed off with dichloromethane. Finally, the brown solid product was dried in vacuum oven at 60 ℃ overnight and collected.

2.3 Physical characterization

X-Ray Diffraction (XRD) test was determined by a D/max-2000 XRD diffractometer (Cu target, λ = 1.50456 Å) from Rigaku Co., Japan. Scanning electron microscope (SEM) measurement was made by supra 55 sapphire FE-SEM of Carl Zeiss, Germany. Transmission electron microscope (TEM) determination was conducted with JEOL’s JEM-2100 F. Brunner-Emmet-Teller (BET) data were obtained by measuring N2 adsorption-disadsorption isotherms at 77 K using Micromeritics ASAP 2020 analyzer. X-ray photoelectron spectroscopy (XPS) spectra were characterized with a thermo escalab 250 energy spectrometer and analyzed with XPSPEAK software. Ultraviolet–visible (UV–Vis) test was conducted with a mini-1240 UV spectrophotometer of Shimadzu Corporation of Japan.

2.4 Electrochemical measurements

2.4.1 Pretreatment of carbon paper and nafion 117 membrane

As the substrate of cathode, carbon paper was ultrasonicated for 10 min in isopropanol solution, and then immersed in concentrated sulfuric acid and hydrogen peroxide solution, respectively, for 5 h at 50 ℃. Finally, it was washed with deionized water to neutral, and then kept in deionized water.

In the case of the Nafion 117 membrane, it was put into the pre-prepared H2O2 solution (5 vol%) and heated in a water bath at 85 ℃ for 2 h. After being cooled to room temperature, it was boiled in deionized water at 55 ℃ for 30 min, and then boiled in 0.5 M sulfuric acid solution for 2 h at 85 ℃. Finally, the boiled Nafion 117 membrane was washed to neutral and stored in deionized water.

2.4.2 Cathode preparation process

2 mg Fe-TPPHZ, 2 mg Keqin black, 0.5 mL isopropanol and 0.5 mL deionized water were mixed together and ultrasonicated for 30 min. Subsequently, 75 µL Nafion solution (5 wt%) was further added and ultrasonicated for 1 h. Finally, the above uniformly dispersed ink was dropped on the pretreated carbon paper (1 cm2) or the rotating disk electrode (RDE, 0.1963 cm2), respectively.

2.4.3 NRR potential selection

Linear sweep voltammetry (LSV) experiments were conducted in Ar- or N2-saturated 0.1 M KOH solution at room temperature with a rotating disk electrode. The LSV curve was obtained at a scanning rate of 5 mV s− 1 with a speed of 1,600 rpm. All applied potentials were converted to the reversible hydrogen electrode (RHE) scale via the following calibration equation: E(vs. RHE) = E(vs. Hg/HgO) + 0.059*PH + 0.098 [35, 36]. Based on the difference in current densities determined in Ar- or N2-saturated electrolytes, the potentials that used for the NRR test reaction was selected.

2.4.4 NRR test

Electrohydrogenation of N2 to NH3 using Fe-TPPHZ as catalyst was made with a three-electrode system in a typical H-type two-compartment cell. In the whole NRR test process, the above pretreated Nafion 117 membrane was completely immersed to separate cathode and anode chamber.

In the assigned three-electrode system, the Fe-TPPHZ coated carbon paper, platinum sheet and saturated KOH filled Hg/HgO electrode act as the working, counter and reference electrodes, respectively. The chronoamperometry experiments at various measured potentials were conducted with 45 mL 0.1 M KOH as electrolytes in both anode and cathode chamber. N2 bubbles were filled into the KOH electrolytes for 30 min at the outset of the experiment, and then were continuously purged during NRR measurements. Moreover, magnetic stirring was performed at a rate of 600 rpm throughout the measurements, and a continuous acid trap with 5 mL 0.05 M H2SO4 was connected with the cathode chamber. All electrochemical measurements were carried out using a CHI 660E electrochemical analyzer (CH instrument, China) and the applied potentials were calibrated to be that versus RHE.

2.4.5 Quantitative detection of ammonia

Ammonia concentration was quantitatively determined via the generally acknowledged indophenol blue method using the UV–Vis spectrophotometry [22]. Firstly, standard NH4Cl solution was utilized to draw the calibration curve of ammonia concentration and absorbance. The calibration process was conducted independently for three times. In brief, pre-dried NH4Cl was dissolved in 0.1 M KOH and diluted to different concentrations. Then, 2 mL NH4Cl solution with desired concentration was sequentially added with 2 mL of 1 M NaOH solution containing 5 wt% sodium salicylate and 5 wt% seignette salt as a colour-producing reagent, 1 mL of 0.05 M iodine pre-calibrated NaClO as an oxidation solution and 0.2 mL of 1 wt% sodium nitroferricyanide as a catalytic reagent. After 1 h oxidation reaction at room temperature, the absorption spectrum of the above solution was measured using UV–Vis spectrophotometer. Based on the relationship of ammonia concentration and absorbance at 654 nm, a fitting curve was obtained, showing a reliable and accurate linear relation (y = 0.336x + 0.172, R2 = 0.999) (Figure S1). After the electrocatalytic nitrogen reduction reaction, the electrolyte was carefully taken out from the cathodic chamber and added with the same colour-producing, oxidation and catalytic reagent as the above NH4Cl solution. In the light of the UV–Vis absorbance at 654 nm, the NH3 concentration in the cathode chamber can be calculated.

In order to determine the NH3 concentration in the acid trap that was connected with the cathode chamber, a similar calibration curve in the 0.1 M H2SO4 media was carried out, which presents a good linear relationship (y = 0.3624x + 0.0554, R2 = 0.999 ) (Figure S2).

2.4.6 Hydrazine (N2H4) detection

The production of N2H4 was assayed using the reported Watt and Chrisp method [37]. Specially, 5.99 g C9H11NO, 30 mL concentrated HCl and 300 mL C2H5OH were mixed together under magnetic stirring and utilized as the color reagent. Then, a series of 5 mL hydrazine dihydrochloride solutions of known concentrations were separately mixed with 5 mL of the above color reagent under magnetic stirring for 20 min at 25 ℃. Subsequently, the calibration curve was acquired by measuring their UV–Vis absorbance at 455 nm, and the linear relation can be displayed as y = 1.061x + 0.011 with a good linear correlation (R2 = 0.999) (Figure S3). After the electrochemical reduction process, 5 mL electrolyte was added into 5 mL color reagent, and then detected with the same procedure. The amount of N2H4 can be ascertained according to the pre-measured calibration curve of the hydrazine dihydrochloride.

2.4.7 Calculation of NH3 yield and faraday efficiency (FE)

The NH3 yield was the sum of produced NH3 in the cathode chamber, anode chamber and acid trap, which was calculated according to the following equation [38]:

$$\text{Yield}_{(NH3)} = (c\times V)/(t \times m)$$
(1)

where c, V, t and m represents the determined NH3 concentration, volume of electrolyte or acid trap, NRR test time and mass of the Fe-TPPHZ catalyst, respectively.

The FE was calculated based on the following equation and the assumption that three electrons were needed to produce one NH3 molecule [38]:

$${\text{FE }} = {\text{ 3}}F \times c \times V/{\text{ }}\left( {M \times Q} \right)$$

where F, c, V, M and Q stands for the Faraday constant, NH3 concentration, volume of electrolyte or acid trap, molecular weight of NH3 and charge quantity at diverse applied potential that comes from the integral of I–t curves.

3 Results and discussions

3.1 Physical characterization

The composition and surface chemical state were investigated by XPS. Figure 1 shows the XPS survey spectrum of Fe-TPPHZ and the corresponding fitted deconvolutions of Fe 2p, N 1s and C 1s. As illustrated, the peaks centered at ca. 283, 399, 530 and 710.08 eV can be ascribed to the Fe, C, N and O elements, respectively (Fig. 1a). For the fitted high-resolution spectra of Fe 2p, the peaks situated at 710 eV (Fe 2p3/2) and 723.3 eV (Fe 2p1/2) can be assigned to Fe2+, and the peaks at 713.1 eV (Fe 2p3/2) and 725.2 eV(Fe 2p1/2) belongs to Fe3+(Fig. 1b). Moreover, the peak integral area ratio of the Fe2+ and Fe3+ in Fe-TPPHZ is 1:1, indicating that the percentage of Fe2+ and Fe3+ is almost the same. The N 1s spectrum shows two peaks at 398.5 and 399.1 eV (Fig. 1c), which is the typical signal of pyridine nitrogen and pyrrole nitrogen. The fine spectrum of C 1s shows that the binding energy peaks located at 284.7, 285.6 and 288 eV are attributed to C–C, C–N and C=O functional groups, respectively (Fig. 1d). The existence of the pyridine and pyrrole nitrogen atoms can realize the strong adsorption of nitrogen molecules and the easy dissociation of their triple bonds [39], introducing important synergistic electrocatalytic performances on the interface [40]. Furthermore, carbon atoms derived from the conjugated TPPHZ molecular contributes to the elevation of electronic conductivity of catalysts.

Fig. 1
figure 1

XPS survey spectrum of Fe-TPPHZ (a) and the corresponding fitted deconvolutions of Fe 2p (b), N 1s (c) and C 1s (d)

Full size image

The morphology and nanostructure of the as-synthesized Fe-TPPHZ was identified by SEM and TEM. Figure 2 shows the SEM and TEM images of the Fe-TPPHZ. As seen from the images, the Fe-TPPHZ presents an aggregated two-dimensional nanosheet structure (Fig. 2c) and their thickness is less than 5 nm (Fig. 2d). In addition, several thin nanosheets stacked together and formed smooth surfaces and rough edges, which helps to enhance the catalytical effect on NRR because that the sheet-like morphology can bring an extended specific surface area and affluent catalytic active sites [41].

Fig. 2
figure 2

SEM (a, b) and TEM (c) pictures of the as-prepared Fe-TPPHZ

Full size image

The crystallinity structure of the samples was analyzed by X-ray diffraction meter. Figure 3 presents XRD patterns of the TPPHZ and as-prepared Fe-TPPHZ. As illustrated, the TPPHZ shows the characterized peak at 10.3o, 26o, 29.1 o, 39.4 o and 43.1 o. After being conjugated with Fe ions, the peak location of the TPPHZ almost remains the same, but their relative peak intensity changed. Furthermore, new peaks at 11.5, 33.2o and 50.3o appeared, indicating the successful formation of Fe-TPPHZ molecules.

Fig. 3
figure 3

XRD patterns of the TPPHZ (a) and as-prepared Fe-TPPHZ (b)

Full size image

Figure 4 displays the BET data of the as-prepared Fe-TPPHZ. The corresponding value is shown in Table 1. The values of the specific area, pore volume and average pore size are 268. 6 m2 g−1, 0.436 cm3 g−1 and 5.3 nm. This high specific area and nanoporous structure help to provide more active sites to adsorb and catalyze nitrogen reduction reaction.

Fig. 4
figure 4

BET data of the as-prepared Fe-TPPHZ

Full size image
Table 1 BET Parameters of the as-prepared Fe-TPPHZ
Full size table

3.2 NRR performances

As proved in literature [42], K+ ion in the electrolyte can closely interrelate with nitrogen molecules, promote their stern layer interlacement, and enhance the aggregation of nitrogen on the surface of catalysts. Meanwhile, alkaline electrolytes are more effective to weaken the interference of HER than acidic medium [11]. Thus, 0.1 M KOH (pH 13) was selected as the electrolyte for all NRR determination. The whole NRR measurement process was carried out at various applied potential in a conventional H-type electrochemical cell with either N2 or Ar bubbled KOH electrolyte on an electrochemical workstation (CHI 660E) at room temperature. In the three-electrode system, the working and reference electrodes were put in the cathode area, and the counter electrode was separated in the anode area by a Nafion 117 membrane. The continually purged nitrogen in KOH electrolyte around the cathode obtained electrons to generate ammonia, and the hydroxyl ions around the anode converted to oxygen and water by losing electrons. In order to eliminate the possible interferences of ammonia or oxynitride in the inputted N2 or Ar gases, they were pre-purified via a series of devices with 0.1 M NaOH, 0.1 M FeSO4 and 5 mM H2SO4 solutions.

Linear sweep voltammograms (LSV) curves were determined to preliminarily evaluate the electroreduction effect of the as-prepared Fe-TPPHZ catalyst on the NRR test. Figure 5 presents the LSV curves of the Fe-TPPHZ in Ar- or N2-saturated KOH electrolyte. The reduction current increase in the Ar-saturated KOH solution comes from the competing HER, which was used to break the strong H–O–H covalent bonds before hydrogen adsorption. An obvious increase in the reduction current density between − 0.3 to − 0.7 V (vs. RHE) was observed in the N2-saturated electrolyte than in the Ar-saturated one, suggesting a high catalytic effect of Fe-TPPHZ on the reduction of nitrogen in this region.

Fig. 5
figure 5

Linear sweep voltammograms of Fe-TPPHZ in Ar- (black line) or N2-saturated (red line) 0.1 M KOH solution at a scan rate of 5 mV s− 1

Full size image

The NRR performances of the Fe-TPPHZ were then elaborately investigated in a three-electrode system using 0.1 M KOH as electrolyte at various potentials with incessant N2 bubbling in the cathode chamber. All the NRR experiments were performed at room temperature. Meanwhile, a glass tube with diluted H2SO4 was placed at the end of the cell as acid trap to fully adsorb escaped NH3 that was caused by N2 blowing. The total NH3 yield was the sum of produced ammonia in the cathode chamber, anode chamber and acid trap, which was quantitatively examined by the universally acknowledged indophenol blue method. The UV-Vis absorption curves of the electrolytes after the NRR tests are illustrated at different given potentials (Figure S4). Figure 6 presented the average mass-normalized NH3 yield rates and corresponding Faradaic efficiencies of the Fe-TPPHZ at different potentials. As illustrated, the highest NH3 yield rate of 29.07 µg h−1 mg−1 was achieved with an outstanding Faradaic efficiency of 11.5% at -0.3 V vs. RHE. With the negative shift of the given potential from − 0.3 to − 0.7 V, the Faradaic efficiency gradually decreases due to the competing reaction of the HER.

Fig. 6
figure 6

The mass-normalized NH3 yield rates (left y axis) and corresponding Faradaic efficiencies (right y axis) of the Fe-TPPHZ at different applied potentials. The error bars show the average value of three measurements

Full size image

In the NRR process, hydrazine (N2H4) was considered as the possible by-product and the systematically detected spectrophotometric results showed that no hydrazine was generated in the Fe-TPPHZ catalyzed system (Figure S5).

Figure 7 displays the mass-normalized NH3 yield rates and corresponding Faradaic efficiencies of the Fe-TPPHZ, pristine Fe and TPPHZ at -0.3 V versus RHE. In the case of the pristine Fe, the NH3 yield rate and Faradaic efficiency at -0.3 V is only 9.77 µg h−1 mg−1 and 0.71%, and the values for the TPPHZ are 5.1 µg h−1 mg−1 and 2.6%, which demonstrated that the Fe-TPPHZ possesses catalytic superiority than Fe and TPPHZ. It can be predicted that the greatly improved NRR performances of the Fe-TPPHZ may originate from the isolating role of the TPPHZ molecular on active Fe sites, which can utilize their conjugated aromatic ring to ensure the independence and homogeneity of the Fe active sites in the course of NRR test [34].

Fig. 7
figure 7

The mass-normalized NH3 yield rates and corresponding Faradaic efficiencies of Fe-TPPHZ, pristine Fe and TPPHZ at − 0.3 V versus RHE in 0.1 M KOH

Full size image

To further assess the catalytical performances of the Fe-TPPHZ, the mass-normalized NH3 yield rate of the Fe-TPPHZ was compared with literature values, as shown in Fig. 8. Clearly, the as-prepared Fe-TPPHZ attained a higher NH3 yield rate than most reported catalysts.

Fig. 8
figure 8

Comparison of the mass-normalized NH3 yield rates of the Fe-TPPHZ with literature values

Full size image

Figure 9 shows the chronoamperometric curves at different applied potentials and the NRR stability test of Fe-TPPHZ at − 0 to −0.3 V versus RHE. The insets show the chronoamperometry test of the Fe-TPPHZ at − 0.3 V for 12 h and the FE change after 8 cycles. Almost no current density change was detected for the Fe-TPPHZ at − 0.3 V during the initial 2 h test and only a small fluctuation in current density was found after the long-time chronoamperometric measurement (Fig. 9A). An ammonia yield rate of 21.86 µg h−1 mg−1 is observed after 100 consecutive cycles with a retention rate of 75.2% (Fig. 9B), indicating a good stability.

Fig. 9
figure 9

A Chronoamperometric curves at different applied potentials and B NRR stability test of Fe-TPPHZ at − 0.3 V versus RHE. The insets show the chronoamperometry test of the Fe-TPPHZ at − 0.3 V for 12 h and the FE change after 8 cycles

Full size image

4 Conclusions

Novel Fe-TPPHZ nanosheets were synthesized via a hydrothermal way and utilized as highly effective catalysts for the electrocatalytic nitrogen reduction under mild conditions. The as-prepared Fe-TPPHZ nanosheets displayed a remarkable catalytic performance, with a high NH3 yield of 29.07 µg h−1 mg−1 and an outstanding Faradaic efficiency of 11.5% at − 0.3 V vs. RHE. Meanwhile, no hydrazine appearance and NH3 yield change was found during the NRR process for 2 h at − 0.3 V vs. RHE, demonstrating a good selectivity and stability. This work developed a new catalyst for electrochemical production of NH3 from N2.