Control of exposed crystal planes of CeO₂ enhances electrocatalytic nitrate reduction

INTRODUCTION

Nitrate pollution poses a major threat to water resources and ecosystems, leading to adverse health effects and ecological imbalances^[1,2]. By utilizing electrocatalytic reduction, nitrate can be converted into ammonia (NH₃) efficiently, which is widely used as a fertilizer and precursor for various industrial processes^[3,4]. This approach offers a promising alternative to traditional methods of NH₃ production, such as the energy-intensive Haber-Bosch process which consumes large amounts of fossil fuels and emits greenhouse gases^[5-7]. The electrocatalytic and thermal catalytic nitrate reduction to produce NH₃ (NRA) provides a greener and sustainable pathway for NH₃ production but also presents from nitrate pollution^[8-10]. In contrast, thermal catalytic NRA requires a hydrogen source, leading to high complexity of the catalytic device. The electrocatalytic NRA operates with a simple experimental device driven by green electric energy, and it looks more promising. Nevertheless, due to the involvement of an eight-electron transfer, the electrocatalytic reaction exhibits complex reaction pathways and intermediates, and the competitive hydrogen evolution reaction (HER) weakens the electrocatalytic efficiency^[11-13]. As a result, there is an urgent need to design electrocatalysts with high efficiency and selectivity for NH₃ synthesis.

Cerium dioxide (CeO₂), as a rare earth metal oxide, possesses a distinctive crystal structure, such as the flexible conversion between Ce⁴⁺ and Ce^3+[14-16]. These unique characteristics contribute to its excellent redox properties and high oxygen mobility and make it highly valuable in various electrochemical fields^[17-22]. Electrocatalytic procedures are interface reactions conducted on the exposed planes of electrocatalysts. Many works have found that catalytic performance is strongly dependent on the exposed crystal planes, because the exposed planes would interact with the reactants/intermediates resulting in different adsorption energy, Gibbs free energy and activity of competitive/side reactions^[23,24]. Therefore, the NRA performance can be enhanced by controlling the exposed crystal planes. Moreover, the effect mechanism of different exposed crystal planes of CeO₂ on the NRA processes remains ambiguous.

Herein, the exposed crystal planes of CeO₂ are controlled by adjusting the quantity of cetyltrimethylammonium bromide (CTAB), and CeO₂ with exposed (111) plane or (220)/(200) planes is prepared. The CeO₂ with exposed (220)/(200) planes presents higher NRA performance than that of CeO₂ with the (111) plane. Compared with the CeO₂ with exposed (111) plane, the NH₃ yield rate (Yield_NH3) and Faraday efficiency (FE) of CeO₂ with exposed (200)/(220) planes increase by 11.9% and 37.7%, respectively. Density functional theory (DFT) is used to reveal the underlying mechanism. This work paves a new pathway for designing high-performance NRA electrocatalysts.

EXPERIMENTAL DETAILS

RESULTS AND DISCUSSION

Figure 1A-C presents the SEM images of the samples; it is evident that the samples predominantly consist of nearly spherical agglomerates. These agglomerates exhibit a wide size distribution, ranging from around 100 to 400 nm. Additionally, with increasing amounts of CTAB, no significant change in the morphology and size of agglomerates indicates that the amount of CTAB has little effect on its morphology and size. The XRD patterns show that the precursor material is cerium formate, which transforms into CeO₂ upon calcination [Figure 1D]. The peaks at 28.6°, 33.2°, 47.6°, 56.5°, 59.2°, 69.4°, 76.9°, 79.3° and 88.6° correspond to (111), (200), (220) (311), (222), (400), (331), (420) and (422) planes of CeO₂ (PDF 34#0394), respectively. Figure 1E and F demonstrates the N₂ adsorption/desorption isotherm curves and pore size distribution curves of samples. Based on the classification of International Union of Pure and Applied Chemistry (IUPAC) adsorption isotherms, those curves correspond to the type III adsorption model indicating the porous nature of samples^[26]. BET specific surface areas of CeO₂-CTAB0.5, CeO₂-CTAB1.0, and CeO₂-CTAB2.0 are 83, 97, and 87 m²/g, respectively, and the average pore diameters calculated by the Barrett-Joyner-Halenda (BJH) method are 16.3, 16.0 and 13.8 nm, respectively. Therefore, adding CTAB does not significantly influence the morphology, size and specific surface area of CeO₂, and its effect on the NRA performance of CeO₂ can be disregarded.

Figure 1. SEM image of (A) CeO₂-CTAB0.5, (B) CeO₂-CTAB1.0 and (C) CeO₂-CTAB2.0. (D) XRD patterns of precursor and calcined product of CeO₂-CTAB 1.0. (E) The N₂ adsorption and desorption isotherm curves. (F) Pore size distribution curves of CeO₂.

To further investigate the microstructure of CeO₂ and assess its influence on performance, TEM testing was conducted [Figure 2]. Figure 2A-C further reveals the aggregation of smaller structures. Interestingly, the ratio of rod-shaped CeO₂ increases significantly with the rising amount of CTAB. For the small granular CeO₂, the observed interplanar spacing is 0.312 nm [Figure 2D], which corresponds to its (111) crystal plane. This crystal plane is considered the most stable low-index plane for CeO₂ and is exposed typically on granular CeO₂ surfaces^[27,28]. For the rod-shaped CeO₂, the observed interplanar spacing is primarily 0.269 nm corresponding to the (200) crystal plane [Figure 2E and F]. Additionally, for clearly identifiable single-crystal particles, fast Fourier transform (FFT) analysis can be applied to analyze the top and bottom exposed planes^[29]. The norms of the vectors depicted in Figure 2E are all 0.269 nm, corresponding to the interplanar spacing of the (200) and (0$$\overline{2}$$0) crystal planes. The angle between the vectors representing the (200) and (0 $$\overline{2}$$0) planes is measured as 90°, matching the interfacial angle between the (200) and (0$$\overline{2}$$0) crystal faces. Therefore, the zone axis perpendicular to the diffraction plane can be identified as the [002] axis, which confirms that the exposed top and bottom surfaces of the rod-shaped CeO₂ [Figure 2E] correspond to the (002) crystal plane. Likewise, it can be deduced that the top and bottom exposed surfaces of the rod-shaped CeO₂ shown in Figure 2F are (202) surfaces. Thus, the primary exposed surfaces of the rod-shaped CeO₂ are mainly (200) and (220), while that for granular CeO₂ is (111). The results align with previous studies^[30,31].

Figure 2. TEM morphology of (A) CeO₂-CTAB0.5, (B) CeO₂-CTAB1.0 and (C) CeO₂-CTAB2.0. HR-TEM of (D) granular CeO₂ and (E and F) rod-shaped CeO₂ in CeO₂-CTAB1.0.

To determine whether the rod-shaped CeO₂ structures are formed during the precipitation or calcination process, TEM testing was also conducted on the cerium formate precursor [Figure 3A]. However, no rod-shaped morphology is observed in the precursor, suggesting that the formation of rod-shaped CeO₂ structures occurs during subsequent calcination rather than inherited by the precipitation, and the presence of a mushy substance on the surface indicates the presence of CTAB. The DSC/TGA curves of the precursor [Figure 3B] reveal that the main decomposition temperature of Ce(HCOO)₃ is approximately 320 °C, and previous weight loss is primarily due to dehydration and CTAB decomposition^[32]. Based on these findings, the calcination process was adjusted (two-hour calcination at 300 °C followed by one-hour calcination at 500 °C). As a result, all the rod-shaped CeO₂ structures disappear [Figure 3C], and the exposed crystal faces of small particles are (111) plane [Figure 3D]. This outcome can be attributed to the complete decomposition of CTAB at 300 °C, which effectively halts its contribution to the subsequent growth of CeO₂. When the CTAB is removed ahead, the growth of CeO₂ would be along the (111) crystal plane with the lowest energy^[33]. With the adsorption of CTAB, the growth of the (111) crystal plane is obstructed, and its growth would follow other directions to form the rod shape.

Figure 3. (A) TEM morphology and (B) TGA/DSC curves of cerium formate precursor. (C) TEM and (D) HRTEM morphology of CeO₂-CTAB2.0 (2-h calcination at 300 °C followed by 1-h calcination at 500 °C).

The NRA performance of prepared catalysts is evaluated. LSV results [Figure 4A] reveal a significant increase in current density at the same potential after the addition of $$\mathrm{NO}_{3}^{-}$$ to the electrolyte, indicating the pronounced NRA performance of CeO₂. In the absence of $$\mathrm{NO}_{3}^{-}$$, CeO₂-CTAB2.0 exhibits a significant decrease in current density at the same potential, indicating minimal competition from HER^[34]. The LSV curves of the catalysts in the Na₂SO₄ electrolyte are shown in Supplementary Figure 1A. When the current density reaches 10 mA, the overpotential of the three catalysts is -1.14, -1.03 and -1.21 V, respectively, and the corresponding Tafel slopes are 471, 536 and 527 mV dec^-1, respectively Supplementary Figure 1B, further confirming their very poor catalytic HER activity. The HER starting potential occurs at approximately -1.0 V (corresponding to a current density of 10 mA cm^-2). Hence, The -1.0 V is selected as the starting potential to ensure high FE and Yield_NH3. Based on the UV-Vis curves of solutions with various $$\mathrm{NH}_{4}^{+}$$ concentrations [Supplementary Figure 2], the obtained calibration curve with a high R² value (0.99905) is presented in Figure 4B. After potentiostatic tests [Supplementary Figures 3-5], the corresponding Yield_NH3 and FE are calculated [Figure 4C-E]. With increasing amounts of CTAB during preparation, both the Yield_NH3 and FE of the obtained CeO₂ show improvement. The maximum Yield_NH3 and FE are 460 μmol h^-1 mg^-1 and 39.1%, respectively. These enhancements can be attributed to the presence of more rod-shaped CeO₂ structures [(220)/(200) exposed planes] in the CeO₂-CTAB2.0 (discussed later). Furthermore, the cycling performance of CeO₂-CTAB2.0 is assessed [Figure 4F]. After five cycles, there is no significant decrease in both Yield_NH3 and FE, indicating its excellent stability. The chronoamperometry curves and Yield_NH3 of various catalysts in electrolytes with and without $$\mathrm{NO}_{3}^{-}$$ are shown in Supplementary Figure 6. It can be seen that when the electrolyte does not contain nitrate, the current is very weak, and the Yield_NH3 is less than 50 umol h^-1 cm^-2. Considering errors in experiments, the Yield_NH3 can be almost negligible. Therefore, it can be inferred that the produced NH₃ is mainly from $$\mathrm{NO}_{3}^{-}$$ of the electrolyte. Based on the CV curves at various scanning rates, the electrochemical active area (ECSA) of CeO₂-CTAB0.5, CeO₂-CTAB1.0 and CeO₂-CTAB2.0 are estimated [Supplementary Figure 7], and the values are 1.75, 3.25 and 2.25 cm². The CeO₂-CTAB1.0 presents the highest ECSA, while its NRA performance is not the highest, indicating that the NRA performance of CeO₂ is majorly determined by exposed crystal planes. The long-term stability test (-1.05 V) of CeO₂-CTAB2.0 is shown in Supplementary Figure 8, and the current density only slightly increases. In the split XPS Ce 3d spectra [Supplementary Figure 9A], u0, u2, u3, v0, v2 and v3 peaks are the spin-orbit splitting peaks of Ce⁴⁺, and u1 and v1 peaks are the signal of Ce³⁺. Before and after testing, the XPS Ce 3d spectra are almost overlapped. In the XPS O 1s spectra [Supplementary Figure 9B], the peaks at 529.4, 530.8, and 532.4 eV represent lattice oxygen, vacancy oxygen, and adsorbed oxygen, respectively. The intensity ratio of vacancy oxygen and lattice oxygen is almost constant before and after testing. The EPR results [Supplementary Figure 9C] also show that the oxygen vacancy concentration remained basically unchanged before and after the performance test^[35]. The above results confirm its catalytic stability.

Figure 4. (A) LSV curves of CeO₂ in a 0.5 M Na₂SO₄ electrolyte with and without $$\mathrm{NO}_{3}^{-}$$. (B) Calibration curve used for estimation of ammonia-N concentration. (C-E) are potential-dependent FE_NH3 and Yield_NH3 over CeO₂-CTAB0.5, CeO₂-CTAB1.0 and CeO₂-CTAB2.0. (F) The cyclic stability test of CeO₂-CTAB2.0 at -1.05 V.

DFT calculations are employed to investigate the effect of exposed crystal planes on NRA. The calculated model of CeO₂ is shown in Supplementary Figure 10. Based on TEM analysis results, 2 × 2 surfaces of CeO₂ (111), (200), and (220) planes are considered for the calculation processes [Supplementary Figure 11]. The adsorption energies of H and $$\mathrm{NO}_{3}^{-}$$ on different exposed surfaces of CeO₂ are initially examined [Figure 5A]. The adsorption for $$\mathrm{NO}_{3}^{-}$$ and H is the first step for the conventional and surface hydrogenation mechanisms, respectively, and the high adsorption energy (E_ad) for $$\mathrm{NO}_{3}^{-}$$ is necessary for the NRA. The CeO₂ (111) plane presents weak adsorption for $$\mathrm{NO}_{3}^{-}$$, and it is hard to conduct NRA reactions. The E_ad for H and $$\mathrm{NO}_{3}^{-}$$ is comparable to the CeO₂ (200) plane. For the (220) plane, the adsorption for $$\mathrm{NO}_{3}^{-}$$ is strong while the adsorption for H is weak. The results indicate that the NRA on the CeO₂ (220) and (200) planes may be conducted by gradually continuous and surface hydrogenation mechanisms, respectively^[36].

Figure 5. (A) Adsorption energy of $$\mathrm{NO}_{3}^{-}$$ and H on CeO₂ (200), (220) and (111) planes. (B) Gibbs free energy diagrams of different intermediates generated during NRA on CeO₂ (200) and CeO₂ (220) planes.

The change of Gibbs free energy throughout the entire NRA reactions is calculated for (200) and (220) planes [Figure 5B] and the configurations are shown in Supplementary Figures 12 and 13. For the CeO₂ with (220) plane, the rate-determining step (RDS) is *NO to *N step with an energy barrier of 3.5 eV. For the CeO₂ with (200) plane, the RDS is *NH to *NH₂ step with an energy barrier of 2.1 eV. For the aspect of free energy, the NRA tends to occur on the CeO₂ (200) plane. The energy barrier (2.1 eV) is close to many reported high-performance NRA electrocatalysts, such as TiO₂^[37], Rh^[38], and NiCo₂O₄^[39]. The improved NRA performance observed in experiments is attributed to the high affinity for $$\mathrm{NO}_{3}^{-}$$ and relatively low energy barrier.

CONCLUSIONS

In this work, the exposed crystal planes of CeO₂ are controlled by adjusting the amount of added CTAB. With the increase of CTAB, the exposed crystal planes of CeO₂ are transferred from (111) to (200)/(220) planes. This transition occurred due to the obstruction of the growth of the (111) crystal plane by the adsorbed CTAB.

The CeO₂ with exposed (220)/(200) planes presents higher NRA performance than that of CeO₂ with the (111) plane. Compared with the CeO₂ with exposed (111) plane, the Yield_NH3 and FE of CeO₂ with exposed (200)/(220) planes increase by 11.9% and 37.7%, respectively. DFT is used to reveal the underlying mechanism. The exposed (111) plane of CeO₂ is hard to adsorb $$\mathrm{NO}_{3}^{-}$$, while both exposed (200)/(220) planes of CeO₂ show high affinity for $$\mathrm{NO}_{3}^{-}$$ and relatively low energy barriers along the NRA pathways, bringing about enhanced NRA performance. This work shows an effective way to improve the catalytic performance of electrocatalysts.