An Interface Heterostructure of NiO and CeO2 for Using Electrolytes of Low-Temperature Solid Oxide Fuel Cells

Interface engineering can be used to tune the properties of heterostructure materials at an atomic level, yielding exceptional final physical properties. In this work, we synthesized a heterostructure of a p-type semiconductor (NiO) and an n-type semiconductor (CeO2) for solid oxide fuel cell electrolytes. The CeO2-NiO heterostructure exhibited high ionic conductivity of 0.2 S cm−1 at 530 °C, which was further improved to 0.29 S cm−1 by the introduction of Na+ ions. When it was applied in the fuel cell, an excellent power density of 571 mW cm−1 was obtained, indicating that the CeO2-NiO heterostructure can provide favorable electrolyte functionality. The prepared CeO2-NiO heterostructures possessed both proton and oxygen ionic conductivities, with oxygen ionic conductivity dominating the fuel cell reaction. Further investigations in terms of electrical conductivity and electrode polarization, a proton and oxygen ionic co-conducting mechanism, and a mechanism for blocking electron transport showed that the reconstruction of the energy band at the interfaces was responsible for the enhanced ionic conductivity and cell power output. This work presents a new methodology and scientific understanding of semiconductor-based heterostructures for advanced ceramic fuel cells.


Introduction
Fuel cells efficiently convert the chemical energy of different fuels (e.g., H 2 , CH 4 ) into electricity, avoiding the limitations of the Carnot cycle. Based on the electrolyte type, fuel cells can be classified in five groups: proton exchange membrane fuel cells, solid oxide fuel cells (SOFCs), molten carbonate fuel cells, alkaline fuel cells, and phosphoric acid fuel cells [1]. SOFCs are often used at high temperatures (700-1000 • C), making them the most promising candidates for clean energy since they do not require precious metal catalysts and their all-solid structure alleviates potential erosion [2]. High operating temperatures of SOFCs provide high ionic conductivity but also yield serious problems. The long-term stability of SOFCs is a great challenge. A traditional anode-supported SOFC may suffer from the apparent agglomeration of Ni particles in the Ni-YSZ anode (Ni-Y 2 O 3 stabilized zirconia), and Sr can readily migrate from the La 0.6 Sr 0.4 Co 0.2 Fe 0.8 O 3-δ cathode to the electrolyte layer, yielding high interface resistance [3,4]. Additionally, it is difficult to monitor electrochemical behavior in a fuel cell at high temperatures [5,6]. To reduce the operating temperature of SOFCs, extensive research has focused on new materials. Zhu et al. reported the fabrication of a semiconductor-ionic fuel cell (SIFC) [7,8]. Composite materials made of perovskite semiconductors (e.g., SrFeO 3 [9], Sr 2 Fe 1.5 Mo 0.5 O x [10], phology of the as-prepared materials were analyzed by scanning electron microscopy (SEM; ZEISS Merlin SEM, Oberkochen, Germany) operating at 15 kV. Transmission electron microscopy (TEM) was performed on a Philips CM12/STEM device with an accelerating voltage of 120 kV. X-ray photoelectron spectroscopy (XPS) data were collected on a Physical Electronics Quantum 2000 device (Al Kα X-ray source) for surface and chemical analyses. The fuel cell performance and electrochemical properties were recorded using an electronic load instrument (IT8511, ITECH Electrical Co., Ltd., Shanghai, China) at 530 • C. The flow rate of H 2 and air were 100 and 150 mL min −1 , respectively, at a pressure of 1 atm. Electrochemical impedance spectroscopy (EIS) was employed to investigate the polarization characteristics of the electrode. The EIS measurements were tested under OCVs using an electrochemical workstation (Gamry Instruments, Reference 3000, Warminster, PA, USA) in a frequency range of 0.1-1.0 MHz. To further analyze the mechanism during the electrode process, we used ZSimpWin software (Version 3.1, Echem software, Leeds, UK) to fit the impedance spectra.

Cell Construction and Measurement
The LiNi 0.8 Co 0.15 Al 0.05 O 2−δ (LNCA; Tianjin Bamo Sci. & Tech. Joint Stock Ltd., Tianjin, China) electrode powder was mixed with terpineol to form a slurry, which was then brushed on one side of the Ni foam and dried in an oven at 120 • C for 1 h to obtain the Ni-LNCA electrode. The as-prepared CeO 2 -NiO and CeO 2 -Na-NiO composites were sandwiched between two Ni-LNCA pieces and pressed under a load of 250 MPa to fabricate a single-cell sample with an effective area of 0.64 cm 2 and a thickness of approximately 2 mm. The as-fabricated single cells were symmetrical structures of Ni-LNCA/CeO 2 -NiO/LNCAL-Ni or Ni-LNCA/CeO 2 -Na-NiO/LNCAL-Ni. The flow rates were controlled at 100-120 mL min −1 for H 2 and 150-200 mL min −1 for air at a pressure of 1 atm. All samples were tested at 530 • C after sintering at 600 • C under air for 0.5 h. Figure 1 shows XRD patterns of the as-prepared CeO 2 -NiO and CeO 2 -Na-NiO samples. The characteristic peaks were assigned to CeO 2 (PDF#34-0394) and NiO (PDF#47-1049), indicating the coexistence of CeO 2 and NiO in the as-prepared CeO 2 -NiO heterostructure composite. In the CeO 2 -Na-NiO sample, the phase at 29.26 • can be well assigned to Na 2 O 2 (PDF#16-0270), indicating that Na can be found in the CeO 2 -Na-NiO sample. Any other peaks originating from a chemical reaction between CeO 2 and NiO did not appear in the XRD pattern. also found that SEM image of electrode layer at higher magnification of 10 Kx which is inset in Figure 2d, indicating that the electrode materials are homogeneous particles.    The morphology and microstructure of the commercial CeO 2 and the as-prepared CeO 2 -NiO and CeO 2 -Na-NiO are shown in Figure 2a-c. The SEM micrographs of the commercial CeO 2 and the prepared CeO 2 -NiO were obtained at a magnification of 50 kX (Figure 2a,b), and a magnification of 100 kX was used for the CeO 2 -Na-NiO sample (Figure 2c). This allowed us to observe nanometer-sized features. The commercial CeO 2 had particles between 200 and 900 nm (Figure 2a). The as-prepared CeO 2 -NiO and CeO 2 -Na-NiO composites had smaller particles of 50 to 200 nm (Figure 2b,c). The small grain size in the nanometer range and enhanced interconnections in the CeO 2 -NiO or CeO 2 -Na-NiO composites may have contributed to their better electrochemical performance [23]. Additionally, the surfaces of the larger CeO 2 particles were coated with smaller NiO particles, forming a significant interfacial area between CeO 2 -NiO (Figure 2b,c). Figure 2d shows the cross-sectional SEM graph of CeO 2 -Na-NiO, demonstrating that the three layers of the device had an electrolyte layer of about 894 µm thick. The SEM image of electrolyte layer at higher magnification of 10 Kx is inset in Figure 2d, showing a dense structure. It can be also found that SEM image of electrode layer at higher magnification of 10 Kx which is inset in Figure 2d, indicating that the electrode materials are homogeneous particles. also found that SEM image of electrode layer at higher magnification of 10 Kx which is inset in Figure 2d, indicating that the electrode materials are homogeneous particles.    The hetero-interfaces between the CeO 2 and NiO were identified using high-resolution transmission electron microscopy (HR-TEM) ( Figure 3). This may be an underpinning factor behind the enhanced ionic conductivity, since the hetero-interfaces between CeO 2 and NiO may provide fast channels for both ion and proton conduction. Well-defined crystalline fringes with lattice spacing of 0.271 nm corresponding to the (200) crystal plane of CeO 2 and 2.41 nm corresponding to the (111) crystal plane of NiO were observed, further shown by the fast Fourier transform (FFT) pattern (inset Figure 2b) which is in line with XRD results. Figure 3d provides the elemental mapping results from the HR-TEM test based on Figure 3c. The distribution of Ce, Ni, and O were clearly observed in the electrolyte materials, as showing in Figure 3e,f, elucidating that the Ce, Ni, and O elements are uniformly distributed over the region. The hetero-interfaces between the CeO2 and NiO were identified using high-resolution transmission electron microscopy (HR-TEM) ( Figure 3). This may be an underpinning factor behind the enhanced ionic conductivity, since the hetero-interfaces between CeO2 and NiO may provide fast channels for both ion and proton conduction. Well-defined crystalline fringes with lattice spacing of 0.271 nm corresponding to the (200) crystal plane of CeO2 and 2.41 nm corresponding to the (111) crystal plane of NiO were observed, further shown by the fast Fourier transform (FFT) pattern (inset Figure 2b) which is in line with XRD results. Figure 3d provides the elemental mapping results from the HR-TEM test based on Figure 3c. The distribution of Ce, Ni, and O were clearly observed in the electrolyte materials, as showing in Figure 3e,f, elucidating that the Ce, Ni, and O elements are uniformly distributed over the region.  The surface chemical state of the as-prepared samples was analyzed using the XPS method ( Figure 4). Ni, Ce, and O were detected in the CeO2-NiO sample, while Ni, Ce, Na, and O were present in the CeO2-Na-NiO sample (Figure 4a). The high-resolution XPS spectrum of Ce 3d is shown in Figure 4b. The vibration of Ce 4+ peaked at a low binding energy and Ce 3+ at higher binding energies. The Ce 4+ peak at ~529.1 eV was assigned to the oxygen atoms in Ce( +4 )-O, while the Ce 3+ peak at 531.2 eV was assigned to oxygendeficient regions at the interface (Figure 4d), which is related to their high ionic conductivities [19].

Crystalline Structure and Morphology
The Ni 2p XPS spectrum of the nanostructured NiO is shown in Figure 4c. The spectrum was divided into two edges due to spin-orbit splitting, namely 2p1/2 (~885-870 eV) and 2p3/2 (~869-845 eV) edges [24]. The main 2p line did not exhibit a significant blue shift compared to that of the corresponding single crystals. In addition, two main satellite structures, at ~1.5 and ~7.0 eV on the high-binding energy side of the main line, were present for both 2p1/2 and 2p3/2 edges, and their positions did not differ significantly from those of NiO single crystals [25]. The most important and striking difference between the XPS line shape of the nanostructured and single-crystal NiO was the observed main line bonding. The surface chemical state of the as-prepared samples was analyzed using the XPS method ( Figure 4). Ni, Ce, and O were detected in the CeO 2 -NiO sample, while Ni, Ce, Na, and O were present in the CeO 2 -Na-NiO sample ( Figure 4a). The high-resolution XPS spectrum of Ce 3d is shown in Figure 4b. The vibration of Ce 4+ peaked at a low binding energy and Ce 3+ at higher binding energies. The Ce 4+ peak at~529.1 eV was assigned to the oxygen atoms in Ce( +4 )-O, while the Ce 3+ peak at 531.2 eV was assigned to oxygen-deficient regions at the interface (Figure 4d), which is related to their high ionic conductivities [19].
The Ni 2p XPS spectrum of the nanostructured NiO is shown in Figure 4c. The spectrum was divided into two edges due to spin-orbit splitting, namely 2p 1/2 (~885-870 eV) and 2p 3/2 (~869-845 eV) edges [24]. The main 2p line did not exhibit a significant blue shift compared to that of the corresponding single crystals. In addition, two main satellite structures, at~1.5 and~7.0 eV on the high-binding energy side of the main line, were present for both 2p 1/2 and 2p 3/2 edges, and their positions did not differ significantly from those of NiO single crystals [25]. The most important and striking difference between the XPS line shape of the nanostructured and single-crystal NiO was the observed main line bonding.
The peak of Na + (1071.40 eV) was also found on the surface since Na + ions can diffuse toward the surface of a composite material. The locally diffused Na + can attract nearby electrons [26], which can reduce electron mobility on the surface, yielding an effective reduction in the internal short-circuit current of the composite electrolyte.  The peak of Na + (1071.40 eV) was also found on the surface since Na + ions can diffuse toward the surface of a composite material. The locally diffused Na + can attract nearby electrons [26], which can reduce electron mobility on the surface, yielding an effective reduction in the internal short-circuit current of the composite electrolyte.

Electrochemical Performance
The current-voltage (I−V) and current-power (I−P) characteristics of the fuel cells with CeO2-NiO and CeO2-Na-NiO interface heterostructures as electrolytes are shown in Figure 5a. A remarkable peak power density of 571 mW cm −2 was obtained for CeO2-Na-NiO, which is significantly higher than the 350 mW cm −2 obtained for CeO2-NiO at 530 °C. As a comparison, pure commercial CeO2 and NiO were also tested under the same conditions. NiO did not exhibit any considerable outputs for practical applications, and OCV values were below 1 V (Figure 6), indicating short-circuit issues. However, the composite of these two semiconductor materials showed enhanced power density. The output of the CeO2-NiO heterostructure fuel cell was 350 mW cm −2 , and the corresponding OCV was 0.92 V, which illustrates minimal electronic short-circuit issues. Na + in the CeO2-NiO composite improved both the power density (571 mW cm −2 ) and OVC (

Electrochemical Performance
The current-voltage (I−V) and current-power (I−P) characteristics of the fuel cells with CeO 2 -NiO and CeO 2 -Na-NiO interface heterostructures as electrolytes are shown in Figure 5a. A remarkable peak power density of 571 mW cm −2 was obtained for CeO 2 -Na-NiO, which is significantly higher than the 350 mW cm −2 obtained for CeO 2 -NiO at 530 • C. As a comparison, pure commercial CeO 2 and NiO were also tested under the same conditions. NiO did not exhibit any considerable outputs for practical applications, and OCV values were below 1 V (Figure 6), indicating short-circuit issues. However, the composite of these two semiconductor materials showed enhanced power density. The output of the CeO 2 -NiO heterostructure fuel cell was 350 mW cm −2 , and the corresponding OCV was 0.92 V, which illustrates minimal electronic short-circuit issues. Na + in the CeO 2 -NiO composite improved both the power density (571 mW cm −2 ) and OVC (1.04 V), indicating that electronic transport in the electrolyte was suppressed. This experimental result was consistent with the XPS analysis.
ogy. Great enhancements in power densities and OCV should originate from high proton and oxygen ionic conductivities in the electrolyte via the interface. This was demonstrated by the lower power densities obtained from the pure CeO2 and NiO and higher power densities obtained from the composite of CeO2-NiO and CeO2-Na-NiO. These as-analyzed interfaces enhanced ionic conductivities, as was also reported for other semiconductor materials [27][28][29].

Electrical Conductivity and Electrode Polarization
To investigate the conductivity mechanism of the fuel cell with the semiconductor composite electrolyte, we performed EIS measurements. Figure 5b shows the electrochemical impedance spectra of the CeO2-NiO and CeO2-Na-NiO samples under fuel cell operating conditions at 530 °C, with an equivalent circuit (R0(R1Q1)(R2Q2)) used to simulate the obtained results. In the equivalent circuit, R0 represents ohmic resistance of the electrolyte, R1 and R2 are polarization resistances, and Q is the constant phase element (CPE). The experimental results were simulated in ZSimpWin software, and the data are summarized in Table 1

Electrical Conductivity and Electrode Polarization
To investigate the conductivity mechanism of the fuel composite electrolyte, we performed EIS measurements. Fi chemical impedance spectra of the CeO2-NiO and CeO2-Na-N operating conditions at 530 °C, with an equivalent circuit (R late the obtained results. In the equivalent circuit, R0 repres electrolyte, R1 and R2 are polarization resistances, and Q is (CPE). The experimental results were simulated in ZSimpWin summarized in Table 1 This phenomenon differed to some extent from the state-of-the-art fuel cell technology. Great enhancements in power densities and OCV should originate from high proton and oxygen ionic conductivities in the electrolyte via the interface. This was demonstrated by the lower power densities obtained from the pure CeO 2 and NiO and higher power densities obtained from the composite of CeO 2 -NiO and CeO 2 -Na-NiO. These as-analyzed interfaces enhanced ionic conductivities, as was also reported for other semiconductor materials [27][28][29].

Electrical Conductivity and Electrode Polarization
To investigate the conductivity mechanism of the fuel cell with the semiconductor composite electrolyte, we performed EIS measurements. Figure 5b shows the electrochemical impedance spectra of the CeO 2 -NiO and CeO 2 -Na-NiO samples under fuel cell operating conditions at 530 • C, with an equivalent circuit (R 0 (R 1 Q 1 )(R 2 Q 2 )) used to simulate the obtained results. In the equivalent circuit, R 0 represents ohmic resistance of the electrolyte, R 1 and R 2 are polarization resistances, and Q is the constant phase element (C PE ). The experimental results were simulated in ZSimpWin software, and the data are summarized in Table 1. The simulated results mainly show three contributions, i.e., one semi-circle with an additional small arc at high frequencies. The high-frequency side reflects the contribution of the grain resistance, the second intermediate frequency area is the contribution of the grain boundary resistance, and the third progress is the polarization resistance reflecting charge transfer behavior at low frequencies [30,31].

Mixed Oxygen-Ion-Proton Conducting Mechanism
As discussed above, ionic conductivity plays a key role in cell performance, while electronic conductivity has an adverse effect. The significantly enhanced power output of the CeO 2 -NiO and CeO 2 -Na-NiO heterostructures could be attributed to a great enhancement in the ionic conductivity due to the interfacial effect since individual CeO 2 or NiO samples did not exhibit good performance. The interface-enhanced ionic conductivity has also been found in other heterostructure composite materials [29,32]. We also found that proton conductivity can occur through CeO 2 -NiO and CeO 2 -Na-NiO electrolyte layers. The ions passed through the perfect bulk lattice, while the proton transport happened through the layer by the interface structure. This is different from the traditional bulk oxygen ion (O 2− ) conduction mechanism due to "proton shuttles", which contribute to much better performance (Figure 7c). NiO is a p-type semiconductor [20], while CeO 2 holds an n-type character [19]. Therefore, a p-n-type contact was constructed at the interface between CeO 2 and NiO. A charge separation mechanism existed at the CeO 2 -NiO interface due to electron transfer from NiO to the CeO 2 (Figure 7b). An electron depletion region formed at the NiO side of the interface and a corresponding electron accumulation region at the CeO 2 side of the interface. Furthermore, the charge separation was additionally enhanced at the operating temperatures of the fuel cells. The positively charged layer in NiO prevented the proton from migrating to the depth of NiO and crossing the interface with CeO 2 due to electrostatic repulsion ( Figure 7b). Consequently, the proton transport was limited to the surface and a shallow layer near the NiO surface region. Due to a weaker H-O interaction and lower activation energy of proton diffusion in NiO, the proton transport was easier in NiO than in CeO 2 . Finally, due to the beneficial blocking effect of the positively charged layer in NiO, the "proton shuttles" performed the transport process in continuous high-conducting regions formed on the SOFC electrolyte membrane (Figure 7c). Martin and Duprez determined the oxygen and hydrogen surface diffusion on the oxide surfaces and pointed out that both oxygen and hydrogen can transport rapidly on the CeO 2 surface [33,34].
conduction may be a dominant ion conduction mechanism for the etched CeO2 electrolyte. The charge carriers of this interfacial conduction phenomenon were determined to contain oxygen ions and protons, as described above. The specific migration mechanism of oxygen ions and protons in CeO2-NiO or CeO2-Na-NiO electrolytes requires further investigation.

Mechanism of Blocking of Electron Transport
The question of how semiconductor interface heterostructures suppress electronic conductivity, which results in high ionic conductivity, needs to be clarified. As reported, the electronic conductivity has both positive and negative impacts on the performance of SOFCs with a semiconductor and ionic composite electrolyte [38]. The appropriate number of electrons in the heterostructure can enhance the triple phase boundary of both anode and cathode functional regions, which can greatly reduce polarization resistance [39]. In contrast, exorbitant electronic conductivity of the composite will induce a short-circuit issue, yielding low OCVs and power outputs. In this work, the semiconducting heterostructure was constructed for a novel electrolyte using a p-type (NiO) and n-type (CeO2) semiconductor. The device with the CeO2-Na-NiO heterostructure exhibited significantly better ionic conductivity and power output, accompanied by high OCVs at low temperatures. The working mechanism was based on a p-n heterojunction in the CeO2-NiO heterostructure membrane, which is a novel aspect of state-of-the-art SOFCs. The heterostructure was observed using the SEM and HR-TEM microscopy and gave insight into the interface conductivity of CeO2-NiO and CeO2-Na-NiO composites.
Generally, when two distinct particles or grains are interconnected, charge redistribution occurs as illustrated in Figure 8, where a desirable p-n heterojunction formed at the interface region between CeO2 and NiO due to different band offsets. This produced a local electric field and a potential gradient at the interfacial region [40]. Additionally, owing to different Fermi levels of CeO2 and NiO, the band inclined at the interface of the CeO2 and NiO heterostructure when two distinct particles or grains were interconnected. The charge transportation occurred from a higher to lower Fermi level to reach an equilibrium state at the interface. The redistribution of charges at the interface between CeO2 and NiO could have been due to the difference in Fermi level positions, valence bands, and bandgap energies. This resulted in the band incline in the CeO2 and NiO heterojunction. Different energy levels and similar Fermi energy levels of CeO2 and NiO induced an Secondly, the NiO-CeO 2 composites exhibited both proton and oxygen ionic conductivities. To prove the existence of proton conductivity in the CeO 2 -NiO and CeO 2 -Na-NiO composites, special cells were fabricated using BZCY in the configuration of Ni-LNCA/BZCY/x/BZCY/LNCA-Ni (x = CeO 2 -NiO, CeO 2 -Na-NiO), which can block the transport of O 2− and e − (Figure 7a). Such special cells allowed only the proton transport through the electrolyte, contributing to the fuel cell output. The proton conductivity of the CeO 2 -NiO and CeO 2 -Na-NiO samples are shown as I-V and I-P characteristics in Figure 7a. The power densities of 148 mW cm −2 and 191 mW cm −2 were determined for CeO 2 -NiO and CeO 2 -Na-NiO with BZCY, respectively. The high current and power outputs confirm the considerable proton conductivity of the as-prepared CeO 2 -NiO and CeO 2 -Na-NiO samples. Bonano [35] and Maria [36] also provided other methods to block the transport of O 2− in the composites.
The proton conductivity (δ iH ) was estimated from the slope of polarization curves in the ohmic polarization region as shown in Table 2.
where δ i is the ionic conductivity, including both proton (δ iH ) and oxygen ionic (δ iO ) conductivities. The δ i and δ iH were estimated from the polarization curve (I−V) of the fuel cells as the linear part of the curve was known since the ohmic resistance was dominated by the electrolyte [37]. According to this method, the proton and oxygen conductivity were calculated as shown in Table 2. The δ iH values of the CeO 2 -NiO and CeO 2 -Na-NiO with BZCY devices represented 37.3% and 29.7% of δ i , respectively. These results are in agreement with the outputs of the CeO 2 -NiO and CeO 2 -Na-NiO with BZCY devices. The outputs of 42.3% and 33.5% contributed to the proton conductivity because BZCY was used to block the O −2 . The little discrepancy between these two datasets is acceptable considering the resistance of BZCY. Table 2. The conductivities of the as-prepared materials at 530 • C.
CeO 2 -NiO 0.204 0.128 0.076 CeO 2 -Na-NiO 0.296 0.208 0.088 As discussed, ion conductivity includes both oxygen and proton contributions. Hence, the partial outputs for CeO 2 -NiO and CeO 2 -Na-NiO of 57.7% and 66.5% must have been caused by oxygen ionic conductivity. The results indicate that the ionic interfacial conduction may be a dominant ion conduction mechanism for the etched CeO 2 electrolyte. The charge carriers of this interfacial conduction phenomenon were determined to contain oxygen ions and protons, as described above. The specific migration mechanism of oxygen ions and protons in CeO 2 -NiO or CeO 2 -Na-NiO electrolytes requires further investigation.

Mechanism of Blocking of Electron Transport
The question of how semiconductor interface heterostructures suppress electronic conductivity, which results in high ionic conductivity, needs to be clarified. As reported, the electronic conductivity has both positive and negative impacts on the performance of SOFCs with a semiconductor and ionic composite electrolyte [38]. The appropriate number of electrons in the heterostructure can enhance the triple phase boundary of both anode and cathode functional regions, which can greatly reduce polarization resistance [39]. In contrast, exorbitant electronic conductivity of the composite will induce a short-circuit issue, yielding low OCVs and power outputs. In this work, the semiconducting heterostructure was constructed for a novel electrolyte using a p-type (NiO) and n-type (CeO 2 ) semiconductor. The device with the CeO 2 -Na-NiO heterostructure exhibited significantly better ionic conductivity and power output, accompanied by high OCVs at low temperatures. The working mechanism was based on a p-n heterojunction in the CeO 2 -NiO heterostructure membrane, which is a novel aspect of state-of-the-art SOFCs. The heterostructure was observed using the SEM and HR-TEM microscopy and gave insight into the interface conductivity of CeO 2 -NiO and CeO 2 -Na-NiO composites.
Generally, when two distinct particles or grains are interconnected, charge redistribution occurs as illustrated in Figure 8, where a desirable p-n heterojunction formed at the interface region between CeO 2 and NiO due to different band offsets. This produced a local electric field and a potential gradient at the interfacial region [40]. Additionally, owing to different Fermi levels of CeO 2 and NiO, the band inclined at the interface of the CeO 2 and NiO heterostructure when two distinct particles or grains were interconnected. The charge transportation occurred from a higher to lower Fermi level to reach an equilibrium state at the interface. The redistribution of charges at the interface between CeO 2 and NiO could have been due to the difference in Fermi level positions, valence bands, and bandgap energies. This resulted in the band incline in the CeO 2 and NiO heterojunction. Different energy levels and similar Fermi energy levels of CeO 2 and NiO induced an adjustment to the conduction band offset (∆Ec) and the valence band offset (∆Ev) at the interface to form potential barriers and a built-in electric field (Figure 8). The principle is similar to that of solar photovoltaic cells. After the built-in electronic field is formed, it can block the electron transport through the electrolyte while charged species (e.g., O 2− or H + ) can be easily moved from one side to the other. According to this new mechanism, it is easy to understand that the CeO2 and NiO heterostructure can indeed favor ionic transport for electrolyte function. As previously reported, the oxygen vacancies at the interface between CeO2 and NiO can be more stable and are easily produced with low formation of energy [41][42][43].
To prove the p-n heterojunction at the interface region between CeO2 and NiO, we prepared and tested a device with a configuration of Ag/CeO2-Na-NiO/Ag. The nonlinear rectification junction characteristic in the measured I-V curves reflected the existence of a built-in heterojunction [44,45], which blocked the electron transport through the device (Figure 9a). It is worth mentioning that the CeO2-Na-NiO sample was run stably for more than 7 h (Figure 9b), indicating that the CeO2-Na-NiO sample could function as an electrolyte for SOFCs with no obvious short-circuit problems during operation. Although stability was obtained in 7 h, long-term durability tests need to be further studied. Unfortunately, after operating for about 7 h, the voltage decreased rapidly, which indicates possible reduction of NiO to Ni. This phenomenon is consistent with the result reported by Liu et al. [20]. We will make further efforts to investigate the degradation mechanism and engineering technology to enhance the stability of the as-prepared device in the future.  The principle is similar to that of solar photovoltaic cells. After the built-in electronic field is formed, it can block the electron transport through the electrolyte while charged species (e.g., O 2− or H + ) can be easily moved from one side to the other. According to this new mechanism, it is easy to understand that the CeO 2 and NiO heterostructure can indeed favor ionic transport for electrolyte function. As previously reported, the oxygen vacancies at the interface between CeO 2 and NiO can be more stable and are easily produced with low formation of energy [41][42][43].
To prove the p-n heterojunction at the interface region between CeO 2 and NiO, we prepared and tested a device with a configuration of Ag/CeO 2 -Na-NiO/Ag. The nonlinear rectification junction characteristic in the measured I-V curves reflected the existence of a built-in heterojunction [44,45], which blocked the electron transport through the device (Figure 9a). It is worth mentioning that the CeO 2 -Na-NiO sample was run stably for more than 7 h (Figure 9b), indicating that the CeO 2 -Na-NiO sample could function as an electrolyte for SOFCs with no obvious short-circuit problems during operation. Although stability was obtained in 7 h, long-term durability tests need to be further studied. Unfortunately, after operating for about 7 h, the voltage decreased rapidly, which indicates possible reduction of NiO to Ni. This phenomenon is consistent with the result reported by Liu et al. [20]. We will make further efforts to investigate the degradation mechanism and engineering technology to enhance the stability of the as-prepared device in the future.  Figure 8. The band structure of CeO2-NiO heterostructure composites.
The principle is similar to that of solar photovoltaic cells. After the built-in electronic field is formed, it can block the electron transport through the electrolyte while charged species (e.g., O 2− or H + ) can be easily moved from one side to the other. According to this new mechanism, it is easy to understand that the CeO2 and NiO heterostructure can indeed favor ionic transport for electrolyte function. As previously reported, the oxygen vacancies at the interface between CeO2 and NiO can be more stable and are easily produced with low formation of energy [41][42][43].
To prove the p-n heterojunction at the interface region between CeO2 and NiO, we prepared and tested a device with a configuration of Ag/CeO2-Na-NiO/Ag. The nonlinear rectification junction characteristic in the measured I-V curves reflected the existence of a built-in heterojunction [44,45], which blocked the electron transport through the device (Figure 9a). It is worth mentioning that the CeO2-Na-NiO sample was run stably for more than 7 h (Figure 9b), indicating that the CeO2-Na-NiO sample could function as an electrolyte for SOFCs with no obvious short-circuit problems during operation. Although stability was obtained in 7 h, long-term durability tests need to be further studied. Unfortunately, after operating for about 7 h, the voltage decreased rapidly, which indicates possible reduction of NiO to Ni. This phenomenon is consistent with the result reported by Liu et al. [20]. We will make further efforts to investigate the degradation mechanism and engineering technology to enhance the stability of the as-prepared device in the future.

Conclusions
In this study, a novel CeO 2 -NiO heterostructure for low-temperature SOFC electrolyte applications was successfully developed. The performance and conductivity of the device with the CeO 2 -NiO heterostructure were significantly enhanced compared to the individual NiO and CeO 2 . The introduction of Na + ions into the composite electrolyte (CeO 2 -Na-NiO) reduced the mobility of electrons on the surface and further improved the overall performance. To establish the experimental descriptions, underlying mechanisms, and functionalities, we employed band alignment to explain the mechanism of ionic conductivity enhancement and the suppression of electronic conductivity. This was proven by the I-V characteristics under biased voltage, which resulted in a semiconductor behaving like a diode, indicating a junction effect in the CeO 2 -NiO fuel cell device. All these findings suggest that the semiconductor interface heterostructure charged reconstruction at the interface between n-type and p-type semiconductor materials as well as in the built-in electric field, playing a key role in the ionic conductivity enhancement and final excellent chemical performance. Therefore, the semiconductor interface heterostructure is a very promising approach for advanced low-temperature SOFCs.

Data Availability Statement:
The data present in this work are available on request from the corresponding author.