Nanotechnologies in ceramic electrochemical cells

Research on nano-size defects in CECs remains scarce making the accurate evaluation of their effects difficult. Even so, some research results about solids provide valuable references. For example, Maier et al. pointed out that when the size decreased at least three effects could be triggered out, i.e. the overall transport property effect, mesoscopic space charge effect, and mesoscopic structural effect.⁷⁰ Drozd et al. employed a modelling method to study the oxygen vacancies in an anatase TiO₂ NP with a size of 1.1 nm, and compared the results to those obtained for a NP model with a twice larger size.⁷¹ It was theoretically supported that the size decrease of anatase NPs greatly facilitated the formation of surface oxygen vacancies and Ti³⁺ ions. These pioneering results indicate the differences in the defect distribution of NPs in comparison with bulk catalysts, and can provide important information for an in-depth understanding of nano-size effects on the materials in CECs.

As discussed above, the oxygen ion diffusion in oxides usually follows the “vacancy mechanism” between different oxygen ion sites, while the transport of protons and electrons complies with the “hopping mechanism”. Fig. 3b and c display the mass transport in a perovskite on the basis of the BSCF lattice model with periodical 2.5 units of the perovskite octahedra in height (about 1 nm). Fig. 3d and e demonstrate the oxygen ion diffusion in the bulk or across the fluorite surface on the basis of 20 mol% Sm doped ceria with three hexahedrons in length (about 1 nm). In these nanograins, the decreased size could probably not change the transport modes of the diffused charges, but would greatly affect the ionic and electronic conductivities through increasing the active site densities, which in turn, boost the mass transfer and energy conversion abilities. On the other hand, as the catalyst size grows (above 100 nm), the bulk lattice will be filled with high stacking, and the total ionic conductivity of the catalyst will be dominated by the bulk conductivity, as comparatively displayed in Fig. 4a and b.

	Fig. 4 (a) Structural BSCF lattice with periodical 2.5 BO6 octahedra along the a axis, dimensionally equalling to 1 nm. (b) BSCF nanocrystalline material with a dimension of about 100 nm along the a axis. (c) A typical p-type semiconductor with a conduction band (CB), valence band (VB) and Femi band (EF) locating near VB. (d) Band bending near the surface in the nano semiconductor. (e) A metal/p-type semiconductor contact with the band bending near the semiconductor surface. (f) A P–N junction contact with the band bending near the contact. (g) Ionic conduction across the surface and inside the lattice, i.e. the boundary conduction and bulk conduction of the ionic conductor. (h) The boundary conduction and bulk conduction in the nano-sized crystalline group, with the same volume as (g) for comparison. Mott–Schottky barrier model for illustrations of: (i) oxygen ion diffusion and (j) proton conduction. Plotted by the authors.

With regard to electronic conduction, the band structure of nano-sized electrode materials may be rearranged, either for p-type or n-type materials. When the grain size turns into the nano scale, high density of defects on the surface would give rise to new features for catalytic reactions. One notable nano-size effect is the formation of a space-charge zone. Obviously, the energy band variation caused by the space-charge effect on the surface will affect the electronic conductivities of the solids through bending the energy band across the surface, as illustrated in Fig. 4c and d. Moreover, the energy band bending of the space-charge zone could be strengthened while different types of grains contact with each other, as sketched in Fig. 4e and f. For instance, the strong contact between a metal particle and a p-type semiconductor pulls the Femi levels of different materials into the same level, which allows the free transport of stimulated charges between them. Another classical model is a P–N junction which is generated between a p-type semiconductor and a n-type semiconductor. The band bending on the interface of the grains would greatly affect the electron transport ability of the hybrid. When being used as the catalyst, the unipolar conductivity of the P–N junction would endow CECs with multiple new fantastic properties.

In addition to the electronic conduction, the oxygen ion conduction in nanocrystalline materials could also be much different from the bulk materials. As we know, for polycrystalline ceramics, the oxygen ion transport can be simply divided into two individual parts: the bulk and surface diffusion. The first part describes the ionic conduction through the sample interior, while the latter one is related to the mass transfer on the grain surface, such as on the electrode which is comprised of O₂ adsorption, dissociation and incorporation of O²⁻ once the material is fed with air.⁷² As described in great many literature studies,^73–75 the activation energy of ionic diffusion along the surface is usually lower than that of the bulk, implying that the oxygen ion diffusion along the surface is more prone to be triggered out. It is reasonable to expect an elevated ionic conductivity once the oxygen ion diffusion is mainly contributed by the activated surface, apart from the grain bulk. As comparatively shown in Fig. 4g and h, once the space was filled with the same volume of ionic conductors, the nanocrystalline sample possesses a larger surface area compared to the bulk sample. With more exposed surfaces to the adsorbed species, high ionic conductivity along the active surfaces may be accomplished. Apart from oxygen ion transport, proton conduction may also be affected by active surfaces. More active sites produce more proton defects. Then, improved proton conduction could be achieved in nano-sized samples. Meanwhile, it was reported that the spacing of interfaces offers a powerful degree of freedom.⁷⁶ Interesting phenomena, such as the effect of the space-charge zone (or the Mott–Schottky depletion layer), may be generated,⁷⁷ and thereby the proton and oxygen ion conduction and the redox reactions on the interfaces may be affected, as sketched in Fig. 4i and j. Therefore, the electrochemical reaction rates may increase extremely with the decrease of the catalyst size.

Consequently, nanotechnologies play tremendous roles in the modification of electrodes and electrolytes. As sketched in Fig. 2, the HOR and HER are two key processes that are used for the assessment of fuel electrode activities in CECs. The H₂ adsorption/desorption, dissociation/association and ionic conduction processes of the catalysts are the main electrochemical steps occurring in a fuel electrode. In addition, the chemical endurance of fuel electrode materials against reducing gas is crucial for maintaining the long-term stability of CECs. For an electrode equipped with a nanocatalyst, the adsorption of H₂ and the dissociation abilities toward protons are able to be strengthened owing to the increased amount of exposed surface areas. Meanwhile, high proton conductivity can be achieved in PCFCs/PCECs as their active surfaces possess high-level oxygen vacancies working as proton transport channels. In OCFCs/OCECs, since water is produced at/fed into fuel electrodes, oxygen ion conduction works as a desirable factor that can influence their catalytic activity. Ni-based cermets, i.e. Ni metal coupled with ceramic materials, are commonly used as fuel electrodes.⁷⁸ Ni serves as the HOR/HER catalytic site while the ceramic component usually acts as the ionic conducting phase. For fuel electrode-supported CECs, the fuel electrode and electrolyte should first be sintered together at high temperature in order to densify the electrolyte membrane. Normally, most NPs would inevitably evolve into large grains on the macro/micro scale at such densification temperature. Nevertheless, once NPs are introduced into the fuel electrode after the densification process, it is reasonable to believe that HOR/HER kinetics will be revitalized. It implies that the nanotechnology is an attractive and efficient pathway toward rational design of advanced fuel electrodes. Then, it is reasonable to understand that the introduction of a mixed ionic and electronic conducting phase into the electrode backbone can increase the surface reaction kinetics and decrease ion diffusion resistance, which has been evidenced by electrochemical impedance spectroscopy (EIS) analysis results.⁷⁹

A typical structure derived from nanotechnology is a nano-sized hybrid electrolyte cell. Here, the hybrid electrolyte is usually composed of two or more nano-scaled components with a heterojunction structure (for example, a P–N junction) in order to suppress internal electronic short circuit. Two different kinds of semiconductors, p-type and n-type semiconductors, are commonly employed to construct a heterojunction. Under real working conditions, orientational electronic conduction is realized. Four advantages of the nano-type electrolyte are as follows. First, since the electrolyte is fabricated at low temperature, plenty of pores connecting with each other could be preserved inside the electrolyte, which drastically benefits the mass transfer of reactants. Second, the ionic conduction of the electrolyte is improved. Unlike dense membranes, the ionic conductivity of the porous electrolyte may be much different. For a conventional high-density electrolyte, the ionic transport resistance along the grain interfaces is usually regarded as a negative factor for the total ionic conduction of the electrolyte. However, when electrolyte materials are processed into small particles (100 nm or below), surficial ionic conduction would be remarkable or even prevail over the bulk conduction due to the absence of surficial defects. As a result, nano-size electrolyte membranes could exhibit higher ionic conductivity than the dense electrolyte membranes with coarse grains.^80,81 Third, the most favorable operating temperature range of nano-electrolyte-based cells is always much lower than that of traditional dense electrolytes. Carbonate-based hybrid CECs are a good example.^82,83 The best operating temperature for this type of cell usually locates at the melting point of the salt, at which the best mobility of the flux along the interface could be achieved. As we know, most fluxes that are implemented in nano-size electrolytes are alkali metal salts, and their melting points are usually below 600 °C. As a result, the operating temperature with superior ionic conductivity is usually set around this temperature. High ionic conductivity, efficient mass transfer, abundant active sites for charged species, etc., effectively accelerate the electrochemical reaction rates for the cell and then drive the fourth benefit-high cell performance at low temperature. Actually, as shown in many literature studies, the cells with nano-materials usually output much higher PPDs, demonstrating positive effects of nanotechnologies on electrolytes.^82,84

In CECs, air electrodes work under an oxidizing atmosphere. As for the air electrodes of OCFCs/ECs, the oxygen ion and electronic conduction is believed to be in high relevance to the electrochemical reactions in air electrodes, while for air electrodes of PCFCs/ECs, triple oxygen ion-proton-electron conducting materials are required to promote the electrode reaction kinetics. In OCFCs/ECs, when using micro/macro-metric particles, the deficient surface area would limit O₂ adsorption and dissociation, in addition to the O²⁻ injection into the lattice, owing to the low oxygen vacancy concentration on the surfaces. In contrast, on NPs, almost all ORR steps could be promoted thanks to the abundant active surface sites, leading to a superior cell performance. The improvement of the oxygen incorporation rate by introduction of NPs supports this viewpoint.⁸⁵ With respect to PCFCs/ECs, the enhanced proton conductivity is helpful to the charge distribution from the TPBs to the whole catalyst surfaces, which in turn drastically accelerates the electrochemical reaction rates, and then high electrode performance could be achieved. Recently, Hong and his co-workers’ EIS analyses of electrodes supported this result.⁸⁶ According to their works, R_s,cat., R_p,cat. and C_p,cat. parameters expressing the activation polarization of the air electrode decreased by approximately one order, 40%, and one order of magnitude, respectively. Apart from the aforementioned parameters, nanometric particles are considered to be in close correlation with the improvement of some key geometric characteristics, such as the tortuosity factors and three-phase boundary length, which are strongly related to the key electrochemical processes.⁸⁷

When investigating the impact of nanotechnologies on cell stability, the accurate definition of NP stability should be firstly provided. In comparison with bulk grains, NPs hold low sizes, more active surfaces, high density of defects, etc. and therefore are more prone to deteriorate thermal and chemical stabilities of devices. High surface energy of NPs can also induce catalyst coarsening at high temperature. Actually, the impact of nanotechnologies on cell stability is still under debate, and there have been at least two different viewpoints in the academic field till now. For instance, to improve the electrode activities, Wang et al. infiltrated Ni–Sm doped CeO_2−δ into a fuel electrode and SmBa_0.5Sr_0.5Co₂O_5+δ into an air electrode.⁸⁹ They ascribed the cell degradation to the high sinterability of NPs which could lead to particle coarsening. In another study, Graves et al. claimed that the degradation mechanism was reversible, and the cell degradation could be eliminated by simply operating the cell under reversible operation cycles between electrolysis and fuel-cell modes.⁹⁰ On the other side, Tong and colleagues’ results evidenced enhanced stability by the infiltration of Gd doped ceria NPs.⁹¹ Besides, Phan and Haes briefly addressed their own points about the NP stability.⁹² They pointed out that the NP stability was highly relative to the targeted size-dependent properties, and can only be available for a finite period of time if all nanostructures are inherently thermodynamically and energetically unfavourable relative to bulk states. Accordingly, stable NPs should at least hold an unchanged aggregation state and core composition and preserve morphology (shape and size) and surface chemistry (original surface potential, chemical identity, etc.) in long-term operation mode. From this point of view, once the NPs exhibit coherent properties in the aggregation state, core composition, morphology and surface chemistry, the cell stabilities would change. Then various situations can be analysed. Herein, for electrodes that are entirely composed of NPs, the stability of heterogeneous multiple types of NPs may be much different. Phase aggregation can be suppressed. In addition, multiple components increase the system chaos and entropy, which in turn benefit the stabilities of chemical composition, morphology, and surface chemistry. A typical example is a one-pot synthesis method. The crystallization of the composite is suppressed, being very helpful for the enhancement of cell endurance. In contrast, an electrode composed of single-phase NPs will be probably reluctant to strong endurance under extremely high temperature and harsh atmosphere conditions, due to the simple component of the electrode which is prone to aggregate in extreme environments. Fortunately, most nanoelectrodes are composed of different phases or of single-phase particles with good spatial distribution, which are beneficial to their stability. In addition, it should be noted that the working mechanism involved in stability of NP catalysts is still under debate. For example, in a report, infiltration of (Sc₂O₃)_0.1(ZrO₂)_0.9 scaffolds with Ni did not improve the cell stability, but brought in a decrease in active TPB density.⁹³ The NP coarsening is considered as the main factor that is responsible for most initial electrochemical degradation of Ni-based SOFCs. In these experiments, over-calcination may be one crucial reason deteriorating the catalytic activities of NPs.

Infiltration is a facile pathway for the preparation of nano-sized fuel electrode particles. Gao et al. proposed a new processing strategy for the fabrication of (La_0.9Sr_0.1)_0.98Ga_0.8Mg_0.2O_3−δ (LSGM0.98), a promising electrolyte in SOFCs, and the as-prepared cell can easily achieve high PPD (>1 Wcm⁻²) at an intermediate temperature range (<650 °C).¹³⁸ In their work, they first cofired the ceramic layer-porous La_0.2Sr_0.8TiO_3−δ (LST) support, porous LSGM and dense LSGM layers, and carried out a process of infiltrating nanometric Ni into porous layers. The novel procedure resulted a low R_p of 0.188 Ω cm² at 650 °C for the cell decorated with an optimized anode functional layer (AFL). Besides, a high PPD value of 1.12 W cm⁻² was achieved at 650 °C. The optimization was attributed to the optimization of electrode materials. This work demonstrated great application potentialities of Ni NP infiltration. In another work, Zhu et al. combined SrFeO_3−δ with Ni NPs for impregnation treatment, and employed it as a fuel electrode for SOECs.¹³⁹ In the viewpoint of the authors, the finely distributed Ni NPs on the substrate enlarged TPBs for efficient electrochemical CO₂ splitting. Electrochemical results showed that the as-tailored SrFeO_3−δ demonstrated better properties, and then performed well in a stability test of 100 hours and 8 redox cycles. Wang et al. infiltrated Ni cocatalysts into a Sr₂Fe_1.5Mo_0.5O_6−δ (SFM)–Ce_0.8Sm_0.2O_1.9 (SDC) fuel electrode to facilitate the electrolysis process for a methane-involved reaction.¹⁴⁰ It was found that, after Ni infiltration, the surface oxygen exchange coefficient greatly increased by about 7 fold and the current density was strongly enhanced over twice at 850 °C. The SFM scaffold was infiltrated by Ni nitrate as well, working as a reversible CEC (R-CEC), as reported by Xu and colleagues (Fig. 6a).¹⁴¹ Compared to a bare SFM-based SOFC, the PPD value was increased from 259 mW cm⁻² to 361 mW cm⁻² in a LSGM-supported cell. When exposed to a CO₂–CO mixture gas, a current density of 0.745 A cm⁻² was recorded at 1.6 V and 800 °C in electrolysis mode. Besides, many valuable works concentrated on the infiltration strategy, and considerable enhancements in cell performance were achieved. Herein, some representative literature studies (in the last 5 years) that are related to the infiltration method are summarized by categories of SOFCs, SOECs and R-CECs and presented in Tables 1–3.

	Fig. 6 (a) Micro-morphology of a Ni infiltrated SFM catalyst. Reproduced with permission from 141. Copyright 2021, Elsevier. (b) A schematic comparison of the pristine and reduced SGNM28–GDC. Reproduced with permission from 142. Copyright 2019, American Chemical Society.

Cell configurations	NP compositions	Sizes	Nanotechnologies	Cell performances (Rp and PPDs)	Notes and ref.
NiO–BaZr0.1Ce0.7Y0.1Yb0.1O3−δ (BZCYYb1711, 65:35 wt%)|BZCYYb1711 + 1 wt% NiO|BaCe0.4Fe0.4Co0.2O3−δ (BCFC)	Ce-Rich orthorhombic phase and Fe-rich cubic phase (air electrode)	—	One-pot synthesis	0.075 Ω cm2 at 700 °C, 0.335 W cm−2 at 700 °C	143
NiO–YSZ|YSZ|SDC|BaCoO3−δ (BCO)–La0.6Sr0.4Co0.2Fe0.8O3−δ (LSCF)	BCO	50–200 nm	Infiltration	∼0.05 Ω cm2 in a symmetrical cell, 0.514 W cm−2 at 700 °C	Cr-Tolerance electrode144
NiO–BaZr0.8Y0.2O3−δ (BZY) (60:40 wt%)|BZY|Sm0.5Sr0.5CoO3−δ (SSC)/PrBaCo2O5+δ (PBCO)–BZY	SSC, PBCO	50–80 nm	Infiltration	0.602 W cm−2, 0.08 Ω cm2 at 600 °C (with SSC), 0.650 W cm−2, 0.07 Ω cm2, at 600 °C (with PBCO)	145
NiO–BaZr0.1Ce0.7Y0.2O3−δ (BZCY172)|BZCY172|(Pr0.9La0.1)2(Ni0.74Cu0.21Nb0.05)O4+δ (PLNCN)–BZCY172	PLNCN	10–20 nm when loading 27.3 wt% PLNCN, 50–200 nm when loading 46.1 wt% PLNCN	Infiltration	0.77 W cm−2 and 0.127 Ω cm2, at 700 °C	146
NiO–BaCe0.5Zr0.35Y0.15O3−δ (BZCY53515, 60:40 wt%)|BZCY53515|BZCY53515–La0.6Sr0.4Co0.2Fe0.8O3−δ (LSCF6428)	BZCY53515 (NPs)-embedded LSCF6428 fiber	Fiber (90–150 nm in diameter) embedded with BZCY53515 NPs	Infiltration	0.537 W cm−2 and 0.181 Ω cm2, at 700 °C	147
NiO–BaZr0.2Ce0.6Y0.1Yb0.1O3−δ (BZCYYb2611, 60:40 wt%)|NiO–BZCYYb2611(AFL)|BZCYYb2611|Pd–PrBa0.5Sr0.5Co1.5Fe0.5O5+δ (PBSCF)	Pd-Deposited PBSCF	15 nm-thick layer for the sample deposited for 3 min	Infiltration	0.42 W cm−2 at 500 °C	Low-temperature operation148
NiO–BZCYYb2611 (60:40 wt%)|NiO–BZCYYb2611 (AFL)|BZCYYb2611|PBSCF	Pd deposited fuel electrode	3 nm (apparently)	ALD deposition	0.34 W cm−2 at 500 °C	Ammonia fuel149
Fe22Cr support|La0.4Sr0.4Fe0.03Ni0.03Ti0.94O3 (LSFNT)–FeCr–ScYSZ infiltrated by NiO–GDC20|ScYSZ|GDC20|(La0.6Sr0.4)0.99CoO3−δ (LSC)	Ni (fuel electrode)	Ni 50 nm	Infiltration	0.650 W cm−2 at 0.7 V and 700 °C, with a fuel utilization of 31%	150
La0.6Sr0.2Cr0.85Ni0.15O3−δ (LSCrN)–Ce0.9Gd0.1O1.95 (GDC10) (80:20 wt%)|NiO–GDC10(60:40 wt%)|GDC10|La0.6Sr0.4Co0.2Fe0.8O3−δ (LSCF)–GDC10 (70:30 wt%)	Exsolved Ni NPs (fuel electrode side)	∼30 nm	Exsolution	0.758 W cm−2 and 0.11 Ω cm2 at 750 °C	In 50% CO2–50% CH4151
NiO–SDC|SDC|Ba0.95(Co0.4Fe0.4Zr0.1Y0.1)0.95Ni0.05O3−δ (BCFZYN)	BCFZYN anchored by NiO NPs	—	Exsolution	1.15 W cm−2 (550 °C), 0.036 Ω cm2	OCFC152
NiO–BZCYYb1711 (60:40 wt%)|BZCYYb1711|Ba0.95(Co0.4Fe0.4Zr0.1Y0.1)0.95Ni0.05O3−δ (BCFZYN)	BCFZYN anchored by NiO NPs	—	Exsolution	0.54 W cm−2 (550 °C) and 0.281 Ω cm2	PCFC152
(La0.6Sr0.4)0.95Fe0.9Mo0.1O3−δ (LSFM)/(La0.6Sr0.4)0.95Fe0.7Ni0.2Mo0.1O3−δ (LSFNM)/(La0.6Sr0.4)0.95Fe0.7Co0.2Mo0.1O3−δ (LSFCM)/(La0.6Sr0.4)0.95Fe0.7Co0.1Ni0.1Mo0.1O3−δ (LSFCNM)|BZCY172|LSCF-SDC	Fe/Ni/Co nano alloys	Fe–Ni alloy: 25–30 nm, Fe–Co alloy: 10–20 nm, Fe–Co–Ni alloy: 20–25 nm	Exsolution	0.258 W cm−2 (LSFCNM, 750 °C), 0.54 Ω cm2 (750 °C)	Symmetric cell, ethane fuel153
NiO–YSZ|YSZ|GDC|Ba0.9K0.1Co0.7Fe0.2Y0.1O3−δ (BKCFY)	BaCoO3−δ (BCO) NPs exsolved from BKCFY	—	Exsolution	0.790 W cm−2, 0.048 Ω cm2 (wet air, 700 °C)	154
NiO–BZCYYb1711(60:40 wt%)|NiO-BZCYYb1711 (AFL)|BZCYYb1711|Ag-doped BCFZY	Ag	<50 nm (apparently)	Exsolution	1.2 W cm−2, 0.06 Ω cm2 (650 °C)	155
(Pr0.5Sr0.5)0.9Fe0.8Ru0.1Nb0.1O3−δ (PSFRN)–Ce0.9Gd0.1O1.95 (GDC10) (50:50 wt%)|La0.8Sr0.2Ga0.83Mg0.17O3−δ|La0.6Sr0.4Co0.2Fe0.8O3−δ (LSCF) –GDC (60:40 wt%)	Fe0.7Ru0.3–FeOx	∼50 nm with the shell size of 2–3 nm	Exsolution	0.683 W cm−2 and 0.034 Ω cm2 (wet H2, 750 °C), 0.374 W cm−2 and 0.198 Ω cm2 at 750 °C and wet C3H8	156
La0.43Ca0.37Ti0.94Ni0.06O3−δ (LCTN)|YSZ|(La0.8Sr0.2)0.95MnO3−δ (LSM)–YSZ (50:50 wt%)	Ni	14 ± 3 nm	Exsolution	1.003 W cm−2 at 900 °C	YSZ supported cell, thermal shock treatment157
NiO–SDC (60:40 wt%)|SDC|BaCo0.8Nb0.1Ta0.1O3−δ (BCNT)|Ba0.95Ag0.05Co0.8Nb0.1Ta0.1O3−δ (BACNT)	Ag and BCNT	BCNT: 100 nm, Ag NPs	PLD + exsolution	∼1 W cm−2 at 600 °C, 0.02 Ω cm2 at 650 °C	158
Sr2Fe1.3Mo0.5Ni0.2O6−δ (SFMNi)–YSZ|YSZ|SFMNi–YSZ	Fe–Ni alloy	—	Exsolution	0.116 W cm−2, 0.93 Ω cm2 at 750 °C, fed with wet H2, 0.1 W cm−2 at 750 °C, fed with wet C3H8	Infiltrating symmetrical cell159
NiO–BCZYYb1711|NiO-BCZYYb1711(AFL)|BCZYYb1711|Pr2Ni0.5Mn0.5O4+δ	PrOx	—	One-pot synthesis	0.65 W cm−2 (700 °C), 0.052 Ω cm2 (700 °C) (symmetrical cell)	160
NiO–BZCYYb1711|BZCYYb1711|Ba(CeCo)0.4(FeZr)0.1O3−δ (BCCFZ) precursor	Rhombohedral and cubic phases	Rhombohedral (23.80 nm) and cubic particles (44.01 nm)	One-pot synthesis	1.054 W cm−2, 0.089 Ω cm2 (650 °C) (symmetrical cell)	161
NiO–BZCY271(60:40 wt%)|BZCY271|Sr4Fe4Co2O13+δ precursor-BZCY271 (70:30 wt%)	A tetragonal perovskite (Sr8Fe8O23+δ, 81 wt%) and a spinel phase (Co3O4, 19 wt%)	—	One-pot synthesis	0.535 W cm−2, 0.677 Ω cm2 (550 °C) (coupled with BZCY271)	162
Anodic aluminium oxide|NiO-GDC|YSZ|Pt	Ni, GDC	—	Magnetron sputtering	0.648 W cm−2 and 0.44 Ω cm2 at 500 °C	Nanofibrous texture163

Cell configurations	NP composition	Sizes	Electrolysis current densities	Faraday efficiencies	Notes and ref.
Pt|GDC|La2NiO4−δ (LNO)–La0.8Sr0.2Co0.8Ni0.2O3−δ (LSCN) (4:96 wt%)	LNO NPs	—	An overpotential of 0.104 V at 0.50 A cm−2 (750 °C)	—	1.2 mm GDC electrolyte supported cell, infiltration164
NiO–YSZ|YSZ|Ce0.9Co0.1O2−δ–La1−xSrxMnO3−δ (LSM)–YSZ (60:40 wt%, loaded by 7.8 wt% Ce0.9Co0.1O2−δ (CDC))	CDC NPs	30 nm CDC	1.26 A cm−2 at 1.3 V at 800 °C	100% (assumed value)	Under 50% humidity, infiltration165
Porous 430L steel support|porous 430L/SSZ|porous SSZ, with SmBa0.5Sr0.5Co2O5+δ (SBSCO) infiltrated into air electrodes and SDC/Ni into fuel electrodes	SDC/Ni, SBSCO NPs	20–100 nm for SDC/Ni, 50 nm for SBSCO	0.73 A cm−2 at 1.3 V in 50% H2O–50% H2, 0.95 A cm−2 at 1.5 V in 90% CO2 – 10% CO, at 650 °C	—	Infiltration166
La0.8Sr0.2Cr0.5Fe0.5O3−δ (LSCrF)–YSZ|YSZ|YSZ–LSCrF, impregnated with Ni–SDC catalysts	Ni–SDC NPs	—	∼0.4 A cm−2 at 0.3 V and 850 °C	>90%	Symmetric cell, CH4–assisted co-electrolysis of H2O and CO2,infiltration167
BaCO3-modified NiO–Zr0.85Y0.15O2−δ (YSZ, 60:40 wt%)|YSZ|La0.8Sr0.2MnO3−δ (LSM)–YSZ (60:40 wt%)	BaCO3 NPs	∼70 nm	0.69 A cm−2 at 1.3 V and 800 °C (0.45 A cm−2, without BaCO3 modification)	—	168
GDC20 infiltrated-Sr2Fe1.5Mo0.5O6−δ (SFM) |Gd0.2Ce0.8O1.9 (GDC20)|YSZ|(La0.75Sr0.25)0.95MnO3−δ (LSM)–YSZ (50:50 wt%)	GDC NPs	GDC: 10–20 nm	0.446 A cm−2 at 1.6 V and 800 °C (12.8 wt% GDC loading)	—	Electrolyte-supported SOEC, infiltration169
SDC-modified 430 stainless steel|YSZ|La0.6Sr0.4Co0.2Fe0.8O3−δ (LSCF) –SDC	SDC NPs	SDC: 20–30 nm	1.38 A cm−2 at 1.5 V, at 800 °C	—	CO2 electrolysis, infiltration170
Ce0.9Mn0.1O2−δ (CMO)-infiltrated (La0.75Sr0.25)0.95(Cr0.5Mn0.5)O3−δ–Ce0.8Gd0.2O1.9 (LSCM–GDC20, 60:40 wt%)|YSZ|(La0.8Sr0.2)MnO3−δ–ScSZ ((Sc2O3)0.10(CeO2)0.01(ZrO2)0.89 (LSM–ScSZ, 60:40 wt%))	CMO	10–15 nm	0.52 A cm−2 at 1.8 V and 800 °C	—	Electrolyte supported cell, infiltration171
Pd–GDC20 infiltrated LSCM–YSZ|YSZ|LSCF–YSZ (50:50 wt%)	Pd–GDC10	50–70 nm	0.364 A cm−2 at 1.5 V and 850 °C in CO2–CO (50:50 vol%). A consumption rate of 2362 µmol cm−2 min−1 at 0.5 V	> 90%	Infiltration172
GDC-infiltrated NiO–YSZ|GDC-infiltrated NiO–YSZ (AFL)|YSZ|GDC|La0.6Sr0.4CoO3−δ (LSC) and Gd,Pr-co-doped CeO2 (GPDC)–GDC	GDC-infiltrated fuel electrode, LSC and GPDC–infiltrated oxygen electrode	GDC: 60 nm at fuel electrode	1.2 A cm−2 at 1.3 V and 750 °C	100% (assumed value)	4 × 4 cm2 cell, infiltration173
Ru-CeO2-infiltrated SFM–SDC (50:50 wt%)|SDC|YSZ|LSM–SDC (50:50 wt%)	Ru and CeO2	10–20 nm	1.82 A cm−2 under 1.5 V and 800 °C in pure CO2	—	CO2 electrolysis, infiltration174
NiO–BCZYYb (60:40 wt%)|BCZYYb1711|NbTi0.4Mn0.1(Ni0.5Cu0.5)0.5O4−δ–BCZYYb1711 (65:35 wt%)	Ni–Cu NPs	∼35 nm	∼30 mA cm−2 under 0.8 V in CO2, at 600 °C	99.8% (current efficiency)	Exsolution175
Sr2Fe1.4Ru0.1Mo0.5O6−δ (SFRuM)–GDC (60:40 wt%)|La0.4Ce0.6O2−δ (LDC)|LSGM|Ba0.5Sr0.5Co0.8Fe0.2O3−δ (BSCF)–GDC (50:50 wt%)	Ru–Fe NPs	<3 nm after manipulation	2.25 A cm−2 at 1.6 V, at 800 °C	>97% (apparently)	Exsolution176
Sr2Fe1.5Mo0.5O6−δ–Ce0.8Sm0.2O1.9 (SFM–SDC, 50:50 wt%)|La0.8Sr0.2Ga0.8Mg0.2O3−δ (LSGM)|Ru (1.5 wt%)–coated SFM–SDC (50:50 wt%)	Ru NPs	∼96 nm	1.06 A cm−2 and 0.38 Ω cm2, at 0.6 V and 850 °C	—	Infiltration177

Cell configurations	Nano compositions	Sizes	Nanotechnologies	Cell performances SOFCs/SOECs	Notes and ref.
La0.3Sr0.6Ni0.1Ti0.9O3−δ (LSNT) fiber–GDC10 (70:30 wt%)|(Sc2O3)0.1(CeO2)0.01(ZrO2)0.89 (SSZ)|La0.2Sr0.8MnO3−δ (LSMO) fiber–GDC (70:30 wt%)	Ni and GDC NPs, mesoporous LSNT (fuel electrode)	30 nm Ni + mesoporous fiber	Electrospinning + in situ exsolution	1.16 W cm−2, 0.27 Ω cm2/1.18 A cm−2 at 1.3 V (800 °C)	178
NiO–Y0.08Zr0.92O2−δ (YSZ)|GDC10|YSZ|GDC|Pr0.9Ag0.1Ba0.5Sr0.5Co2O5+δ (Ag–PBSC) nanofibers	PBSC nanofibers, Ag NPs (air electrode)	∼20 nm (Ag)	Electrospinning + in situ exsolution	0.06 Ω cm2, 0.5 W cm−2/0.65 A cm−2, at 700 °C	179
NiO–YSZ|YSZ|La0.8Sr0.2Co0.8Ni0.2O3−δ–Ce0.9Gd0.1O3−δ (LSCN–GDC10, 30:70 wt%)	LSCN NPs	50 nm (LSCN)	Infiltration	0.1 Ω cm2, 1.336 W cm−2/2.34 A cm−2, at 1.6 V and 800 °C	180
NiO–BaZr0.1Ce0.7Y0.1Yb0.1O3−δ (BZCYYb1711) (60:40 wt%)|AFL|BZCYYb1711|BaCoO3−δ (BCO, 6 wt%) deposited La0.6Sr0.4Co0.2Fe0.8O3−δ (LSCF)	BCO NPs	∼50 nm	Infiltration	1.16 W cm−2 and 0.16 Ω cm2 at 600 °C/1.8 A cm−2, at 1.3 V and 600 °C	∼10 µm-thick electrolyte181
(LaSr)0.9Fe0.9Cu0.1O4−δ (LSFCu)–Ba(Zr0.1Ce0.7Y0.2)O3−δ (BZCY172) (70:30 wt%)|BZCY|LSFCu–BZCY172 (70:30 wt%)	Cu NPs	∼50 nm (Cu)	Exsolution	0.573 W cm−2, 0.19 in air and 0.36 Ω cm2 in 5 vol% H2 at 800 °C/1.02 A cm−2, at 1.2 V and 800 °C	Symmetrical cell with 211 µm electolyte182
NiO–YSZ|YSZ|GDC|Pr2Ni0.8Cu0.2O4+δ (PNCO)–coated La0.6Sr0.4Co0.2Fe0.8O3−δ (LSCF)	PNCO nano film (air electrode)	< 10 nm–thick film (apparently)	Infiltration	1.358 W cm−2 and at 750 °C/1.953 A cm−2 at 1.5 V, 0.107 Ω cm2, with 50% H2O at 750 °C	183
NiO–YSZ|Ni–YSZ (AFL)|YSZ|Ni(Mn1/3Cr2/3)2O4 (NMC) infiltrated–Gd0.1Ce0.9O2−δ (GDC10)	Ni, NMC and GDC10	—	Ni: exsolution, NMC and GDC10: Infiltration	1.293 W cm−2 at 800 °C in a fuel electrode-supported cell/2.32 A cm−2 at 2 V in a symmetrical cell at 850 °C	Fast recovery in composition184
CeO2-infiltrated NiO–YSZ (66:34 wt%)|NiO–Fe2O3 (90:10 wt%)|Ce0.6Mn0.3Fe0.1O2−δ (CMF)-TiO2-added Ce0.6La0.4O2−δ (Ti-LDC, 0.5:99.5 wt%)|LSGM|SSC	CeO2-infiltrated fuel electrode	—	Infiltration	0.95 W cm−2 at 600 °C/1.07 A cm−2 at 1.6 V fed with 20% H2O–30% H2–50% Ar, at 600 °C	185
NiO–YSZ (60:40 wt%)|YSZ|SSC/SDC20-infiltrated LSM	SDC and SSC	SDC: 10 nm, SSC: 40–50 nm	Infiltration	1.205 W cm−2 and 0.08 Ω cm2, at 800 °C/1.62 A cm−2 at 1.5 V with 50% H2O–50% H2 + Ar, at 800 °C, for three-times infiltrated cell	186
NiO–YSZ|YSZ|infiltrated LSM–YSZ (50:50 wt% for the functional layer and 80:20 wt% for current collection layer)	CeOx, PrOx, MnOx	20–100 nm	Infiltration	0.91 W cm−2 at 800 °C and 0.018 Ω cm2, at 850 °C (PrOx modification), 1 A cm−2 at 800 °C and 1.3 V (MnOx infiltration)	187
NiO–BZCYYb1711(60:40 wt%)|BZCYYb1711|Ba0.95(Co0.4Fe0.4Zr0.1Y0.1)0.95Ni0.05O3−δ (BCFZYN)	NiO NPs	—	Sol–gel combustion, exsolution	0.936 W cm−2/1.267 A cm−2 at 1.3 V at 600 °C	188
NiO–YSZ(50:50 wt%)|NiO–YSZ(AFL)|YSZ|GDC10|GDC10 decorated PrBa0.8Ca0.2Co2O5+δ (PBCC)	GDC	70 nm	Sol–gel decoration assembly	1.74 W cm−2 at 750 °C/1.77 A cm−2 at 1.3 V and 750 °C	189
NiO–YSZ|Ni–YSZ(AFL)|YSZ|PdO and ZrO2-infiltrated (La0.8Sr0.2)0.95MnO3−δ–YSZ (LSM–YSZ) (50:50 wt%)	PdO and ZrO2	35 nm	Infiltration	1.114 W cm−2/2.322 A cm−2 under 2.0 V (750 °C)	190

One issue remains that the bonding strength of infiltrated NPs with a substrate/scaffold is commonly not strong enough under most conditions, which will suppress the ion exchange between the NPs and the substrate/scaffold. In this regard, Hui et al. used a particular experimental procedure. They used Ni(NO₃)₂ solution in ethanol for infiltration on an A-site deficient La_xSr_1−3x/2TiO_3−δ (LST) perovskite.¹⁹¹ Optimization of the sintering process for Ni infiltration was carried out to achieve appropriate distribution. Moreover, TG analyses indicated that Ni NPs might dissolve into the lattice due to A-site deficiency of perovskite oxide, resulting in strong bonding interfaces between the NPs and the substrate.

Since the fuel electrode works under a reducing atmosphere, it is reasonable to exsolve Ni metal ions from Ni-containing materials through a reduction reaction. It is known as the in situ exsolution method. By altering the Ni content amount in the parent precursor, exsolved Ni metal NPs with different types of morphologies and various sizes will accumulate over the parent surfaces, anchoring on the grains. Kim evaluated the exsolution process of layered perovskite SrGdNi_xMn_1−xO_4±δ (x = 0.2, 0.5, and 0.8) oxides.¹⁴² The H₂ temperature-programmed reduction (H₂-TPR) measurement revealed that considerable exsolution of Ni NPs was observed above 650 °C in SrGdNi_0.2Mn_0.8O_4±δ (SGNM28) when fed with H₂. After being tested in 10 redox cycles at 750 °C, the electrode ASR of the SGNM28 electrode increased 0.027 Ω cm² per cycle in H₂, corresponding to a degradation rate of about 1.78%. An LSGM-supported SOFC coupled with the tailored electrode exhibited a PPD of 1.26 W cm⁻² at 850 °C. Metallic Ni NPs with a diameter of about 25 nm were found to be homogeneously distributed on the SGNM28 surface. The authors recognized that the exsolved NPs significantly enhanced the HOR activity of SGNM28 through accelerating the electro-oxidation reaction at TPBs to surface-absorbed protons, as illustrated in Fig. 6b. Myung et al. illustrated the generation and growth of metal NPs on an oxide electrode with a chemical composition of La_0.43Ca_0.37Ti_0.94Ni_0.06O_3−δ, by electrochemical poling of a cell at 2 V.¹⁹² Very excitingly, it only took a few seconds. The Ni-decorated electrode exhibited excellent performances both in fuel and electrolysis modes. The as-prepared fuel cell delivered a power output of 2 W cm⁻² in humidified H₂, and a current density of 2.75 A cm⁻² at 1.3 V in 50% H₂O/N₂, at 900 °C. Sun from Prof. Luo's research group reported that the exsolution reaction could also be achieved in a Ni-doped La_0.7Sr_0.3CrO₃ (LSCN) fuel electrode.¹⁹³ As shown in Fig. 7a, they calculated the free energy variations of the oxides composed of different metal ions. It was found that Ni, rather than La, Sr and Cr, could be more reducible in a perovskite lattice considering its relatively negative Gibbs free energy (−46.47 kJ mol⁻¹), while the other metal ions held positive values at the same temperature. The non-stoichiometry with partial A-site deficiency in the perovskite greatly increased the mobile oxygen vacancy density, and strongly enhanced the exsolution ability of B-site Ni ions, improving both the electronic conductivity and electrode catalytic activity. When being tested in a sour fuel, a PPD of 460 mW cm⁻² in 5 thousand ppm H₂S–H₂ was obtained, while the PPD of a cell with stoichiometric LSCN only reached 135 mW cm⁻². Duan et al. implemented a solid-state reaction method by thoroughly mixing NiO and the other raw metal oxides/carbonate together, and finally obtained a BaZr_0.1Ce_0.7Y_0.1Yb_0.1O_3−δ (BCZYYb1711) and NiO composite fuel electrode.¹⁴ According to their results, exsolved Ni NPs with a size of less than 100 nm were successfully obtained. The exsolved Ni NPs were found to be beneficial to the HER process in a PCEC and the HOR process in a PCFC, as evidenced in their experimental results.

	Fig. 7 (a) Calculated values of reduction of Gibbs free energies of La, Sr, Ni, Cr oxides for the perovskite-type LSCN-15 (15 mol% Ni) sample at 800 °C. Reproduced with permission from ref. 193. Copyright 2015, the Royal Society of Chemistry. (b) In situ neutron powder diffraction results at different sintering temperatures in the Ni-doped Pr0.5Ba0.5MnO3−δ perovskite. The decrease of the perovskite (*) and increase of Ni peaks revealed the Ni exsolution reaction during the sintering process. Reproduced with permission from ref. 194. Copyright 2020, the Royal Society of Chemistry. (c) An AFM image of the native surface in the La0.4Sr0.4Ni0.03Ti0.97O3−δ specimen after being reduced. (d) An atomic scale model for the presentation of the surficial structure of terraces in (d). Reproduced with permission from ref. 195. Copyright 2015, Springer Nature.

Then many literature studies evidenced that the exsolution technology is an inspirational composition engineering approach. Bahout et al. presented a complex phase evolution of Ni-incorporated Pr_0.5Ba_0.5MnO_3−δ, i.e. (PrBa)_0.975Mn_1.95Ni_0.05O_6−δ.¹⁹⁴ Microstructural characterization demonstrated that the specimen annealed in air at high temperature was actually composed of two different perovskites: orthorhombic Pr_0.65Ba_0.35Mn_0.975Ni_0.025O_3−δ and 2H-hexagonal BaMnO_3−δ with a weight ratio of 3:1. While being annealed at humid hydrogen, MnO escaped out at about 500 °C, and the orthorhombic oxide transformed into a tetragonal phase, then into a cubic phase at ∼665 °C. At 900 °C, Ni metal particles were detected to be evolved out (Fig. 7b), leaving two perovskite phases being transformed into a layered double perovskite-type oxide, PrBaMn₂O_5+δ, without any Ni element. The EIS results as obtained in a reducing atmosphere at 850 °C demonstrated a polarization resistance value of 0.135 Ω cm² for the Ni-exsolved electrode. Managutti et al. synthesized several Ni-doped oxides with a chemical composition of (Pr_0.5Ba_0.5)_1−x/2Mn_1−x/2Ni_x/2O_3−δ (x = 0, 0.05, 0.1, and 0.2, denoted as PBMN_x) for catalysis application assessment.¹⁹⁶ Phase detection and micromorphology analyses displayed that Ni metal could be exsolved around 875 °C when being subjected to a H₂ atmosphere. The formation of an A-site deficiency in Ni-containing samples was found to be accountable for the exsolution process. When being equipped in a symmetrical cell, the lowest ASR was recorded as about 0.64 Ω cm² at 850 °C for PBMN0.2 in wet 5% H₂/N₂, owing to the high level of the Ni content in PBMNx series. The authors then investigated the dry reforming of methane using Ni-exsolved PBMN0.2, achieving the CH₄ and CO₂ conversion rates of 11 and 32%, respectively, and a H₂ production of 37% at 850 °C. Concurrently, a low level of carbon deposition of 0.017 g gcat⁻¹ h⁻¹ was yielded.

It is frequently recognized that the exsolved particles maintain strong attachment to parent substrates as compared to the coated catalysts via an infiltration strategy. Neagu et al. reported a directional research work on Ni doped La_0.4Sr_0.4TiO_3−δ.¹⁹⁵ They found that, unlike infiltration, the exsolved analogues (Ni) were deeply socketed into the parent material (Fig. 7c and d), implying a stronger metal–oxide interface, which contributed a lot to cell stability and a remarkable decrease of hydrocarbon coking. This work revealed the surface effects and defect interactions on the exsolution process and was valuable for facile realization of perovskites with exsolved NPs for multiple functionalities. More importantly, Kousi et al. proposed a concept that La_0.8Ce_0.1Ni_0.4Ti_0.6O_3−δ endured controllable growth of metallic NPs both in a perovskite oxide lattice and on the surface while appropriately tailoring the preparation procedures (Fig. 8a).¹¹²Fig. 8b and c display the methane-involved conversion mechanism indicating the main stages associated with the redox transformations. It was displayed that, apart from the surficial NPs, the completely embedded NPs in the bulk could exchange oxygen ions with the methane outside, enabling redox conversion to syngas with excellent selectivity and cyclability when surficial NPs were present. La_0.65Sr_0.3Cr_0.85Ni_0.15O_3−δ (LSCrN) was synthesized, which was characterized by surficial exsolved Ni NPs, as reported by Amaya–Dueñas et al. The authors employed it as a fuel electrode on a 5 × 5 cm² electrolyte-supported cell.¹⁹⁷ Ni NPs were clearly observed on LSCrN. Reversible operation (SOFC/SOEC) at 860 °C with a 50% H₂O/50% H₂ gas mixture was tested and the co-electrolysis operation was also conducted for over 950 h. A voltage degradation of lower than 3.5 mV per 1000 h was observed. In-depth results indicated that the exsolved NPs were stable in morphology when the electrode was performed isothermally. In these reports, the working temperature of the cell was normally lower than the temperature at which the exsolution process took place in order to maintain catalytic stability. Moreover, a co-catalytic layer at a fuel electrode can be modified by nanotechnologies. Bae et al. used a NiAl₂O₄ spinel catalyst as the precursor for the synthesis of Ni NPs, which were successfully used for dry reforming of methane at a fuel electrode.¹⁹⁸ It was observed that Ni NPs on the order of 10 nm anchored on the Al₂O₃ support after the reduction process. The results showed that the spinel-derived Ni displayed a single rate-limiting factor of diffusion limitation at all temperature ranges avoiding suppression of surface kinetics, and demonstrated higher stability than a conventional Ni/Al₂O₃ catalyst, which was attributed to the enhanced interactions between the active Ni NPs and the support material. With appropriate design of parent compositions, more oxides, such as La_0.95Fe_0.80Ni_0.05Ti_0.15O_3−δ,¹⁹⁹ Sr_0.8Ti_0.85Ni_0.1Si_0.05O_3−δ, Sr_0.8Ti_0.85Ni_0.1P_0.05O_3−δ,²⁰⁰etc. were successfully used for Ni exsolution.

	Fig. 8 (a) Cross-section SEM view after exsolution indicating the generation of surface and bulk NPs. (b) Curves of CH4 consumption and H2, CO2, CO products. (c) A probable illustration of methane conversion pathways. Reproduced with permission from ref. 112. Copyright 2020, John Wiley and Sons. (d) Atomic scale mechanistic insight into particle exsolution. (e) Plot of the number of Ni atoms involved in a particle as a function of reduction time. Reproduced with permission from ref. 201. Copyright 2019, American Chemical Society.

High-temperature treatment is also an effective way to realize the exsolution process. Tan et al. revealed a novel thermally driven exsolution phenomenon.²⁰² In their experiments, Ni, Gd co-doped ceria was implemented. At high temperature, Ni dopants thermally exsolved, forming NPs with good attachment to the host surface, which indicated a new pathway in assembling exsolved Ni nanocatalysts. A cell with thermally exsolved Ni nanocatalysts from the 5 mol% Ni-doped specimen displayed a comparable R_p value to that of the traditional mechanically mixed Ni-GDC (Gd doped ceria), together with an improvement in TPB density, which were proved to be appropriate candidates as electrodes for SOFCs at low temperature.

Understanding the exsolution mechanism of NPs through visible investigation methods is meaningful for efficiently controlling NP growth, tailoring catalyst activity and enhancing cell stability. Neagu et al. used a high-resolution transmission electron microscope (TEM) to investigate the exsolution mechanism of individual NPs on a perovskite surface.²⁰¹ As schematically displayed in Fig. 8d and e, it was revealed that the socketed, strain-inducing interface endowed exsolved NPs with high stability and reactivity. These results also provided critical information for building intriguing heterostructures and controlling the morphology of the exsolved NPs, which could be achieved by changing the gas atmosphere.

Some fantastic but efficient technologies were also reported toward modification of the size and morphology of Ni-based fuel electrodes. For example, Błaszczak et al. developed a one-pot synthesis route in the preparation of porous YSZ and Ni–YSZ particles.²⁰³ The author claimed that a wormhole-like framework with the generation of Ni NPs was realized. The decrease of the particle sizes in the electrode, high homogeneity, a well developed TPB, and an improved interfacial interaction between Ni and YSZ were regarded as the favourable factors that were very helpful for improving the fuel electrode properties.

	Fig. 9 (a) Thermogravimetric analysis results of PBMCo in the presence of H2–N2 (10–90) composite gas. (b) In situ TEM observations of PBMCo fed with 0.5 Pa of H2 with the temperature changing from 770 (left) to 850 °C (right) with an interval of 40 °C. Reproduced with permission from ref. 205. Copyright 2016, American Chemical Society. (c) Current–voltage curves of a SOEC with Co-decorated LSCM for CO2 electrolysis. Reproduced with permission from ref. 206. Copyright 2019, Elsevier. (d) Schematic of the key reaction processes occurred on a Co-modified SFCoM surface. Replotted by the authors for illustration. (e) Comparison of RE1 (high-frequency resistance corresponding to charge transfer process) and RE2 (low frequency resistance associated with the oxygen surface process) for e-SANC and SNC0.95 specimens, tested between 550 and 650 °C. Reproduced with permission from ref. 208. Copyright 2016, American Chemical Society. (f) DFT calculation results for elucidating the co-segregation energies of Pr (grey), Ba (green), Mn (dark blue), dopants (Mn, Co, Ni and Fe, purple) and O (red) atoms. (g) A comparison of the co-segregation energies of the doped materials. Reproduced with permission from ref. 209. Copyright 2017, Springer Nature.

Fe-based NPs are a promising candidate for the construction of surficial active sites for diverse electrochemical reactions. Tsekouras and the co-worker from Prof. Irvine's group made a step-change in the exsolution of perovskite-type titanates with B-site substitution and A-site cation deficiency.²¹⁰ The chemical formula of the compound designed in their research was La_0.4Sr_0.4M_xTi_1−xO_3−δ (M was Fe or Ni, and x was 0.06). It was then employed as a SOEC fuel electrode for steam electrolysis at 900 °C. The ingenious design of A-site deficiency offered sufficient driving force for accurate exsolution of B-site cations, forming metallic NPs. Fe NPs were found in a La_0.4Sr_0.4Ti_0.94Fe_0.06O_3−δ sample, implying the achievement of a successful exsolution reaction from the parent perovskite. The authors partially ascribed the exsolution reaction to the instability of the host lattice related to the change of vacancies. The presence of active Fe NPs and inherent oxygen vacancy concentrations notably reduced the activation barrier to steam splitting in comparison with the unmodified one. In another report, Sun et al. made an experimental comparison between La_0.7−xSr_0.3Cr_1−yFe_yO_3−δ (x = 0, y = 0 and x = 0, y = 0.15) and A-site deficient La_0.6Sr_0.3Cr_0.85Fe_0.15O_3−δ.²¹¹ The outcomes also demonstrated that the presence of A-site deficiency facilitated the in situ exsolution process of transition metal ions. The exsolved Fe NPs on the A-site deficiency specimen provided extra active sites for the fuel oxidation reaction, which was regarded as the primary reason for the electrochemical activity enhancement.

A Sr_0.95Ag_0.05Co_0.9Nb_0.1O_3−δ (SANC) perovskite oxide was prepared by Zhu and colleagues.²⁰⁸ Through a reduction process, nano-scale Ag particles with a size of 5–10 nm were found anchoring on the surface of the host material (labelled as e-SANC). As provided in Fig. 9e, an e-SANC electrode was demonstrated to be more active for the ORR, yielding a relatively low ASR (approximately 0.214 Ω cm² at 500 °C). An OCFC with a decorated fuel electrode achieved an extremely high PPD of 1.116 W cm⁻² at 500 °C and good sustainability for more than 100 h at a current density of 0.625 A cm⁻². The enhanced activity and stability were attributed to the elevated oxygen surface exchange kinetics and successful construction of an Ag NP-anchored Sr_0.95Nb_0.1Co_0.9O_3−δ composite. Additionally, the chemical tolerance against CO₂ was found to be improved too. Kosaka et al. synthesized BaCe_0.8Y_0.1Ru_0.1O_3−δ, and then annealed the perovskite in a reducing atmosphere.²¹² 1–10 nm Ru NPs were obtained. In this report, the reduction temperature showed high correlation with the size of the exsolved particles, leading to enhanced ammonia formation capability. Herein, more representative references are collected, as presented in Tables 1–3.

A unified exsolution mechanism remains the open question till now. Several reports provided interesting and valuable references to unveiling this smart chemical process. Kwon et al. provided their own conclusion for different trends in the exsolution of some typical transition metals, including Mn, Co, Ni and Fe, while the exsolution reaction was evaluated in layered PrBaMn₂O_5+δ oxide on the basis of calculations, as displayed in Fig. 9f and g.²⁰⁹ The calculation results together with the TEM observations presented that Mn, Co and Ni could be easily exsolved from the lattice which was highly related to the transition metal–perovskite reducibility. Importantly, it was theoretically revealed that the co-segregation of B-site dopants and oxygen vacancies acted as a crucial factor in the exsolution process, as analysed by the DFT calculations.

	Fig. 10 (a) SEM image of STFN after the cell test. (b) TEM observation of an exsolved Fe–Ni NP on STFN. (c) Polarization resistances of STF and STFN-based cells in comparison with the cell equipped with a traditional Ni-YSZ electrode. Reproduced with permission from ref. 214. Copyright 2018, Elsevier. HR-TEM observations of the surficial regions of STFN0 (d) and STFN5 (e) with the generation of a surficial exsolution reaction. Reproduced with permission from ref. 215. Copyright 2019, Elsevier.

Nanoalloys can also be implemented as active sites for a fuel reformer, for instance, carbon removal. Bkour et al. established a Ni–Mo decorated YSZ functional layer for reforming isooctane with a less degree of the catalytic coking effect (Fig. 11a).²¹⁷ In this research, an aqueous solution containing Ni and Mo precursors was used and mixed with the YSZ powder. Finally, a Ni–Mo nanoalloy was prepared through a reactive sintering process by drying the mixture overnight at 100 °C, annealing at 500 °C and then reducing at 750 °C in a stream of 50% H₂/He. Yao et al. reported a Ni–La_0.8Sr_0.2FeO_3−δ (LSF) composite that could be used for generation of a Ni–Fe nanoalloy.²¹⁸ In this work, the authors combined impregnation and exsolution processes together. The average sizes of the as-prepared Ni–Fe NPs were controlled at 20–50 nm. A cell with 10 mol.% Ni impregnated Ni-LSF as a fuel electrode achieved a PPD of 550 mW cm⁻² at 700 °C when being ehavi with syngas. The authors claimed that the strong interaction between the Ni–Fe nanoalloy and substrate should be responsible for efficient suppression of carbon deposition. Zhao et al. developed an active Cu–Fe nano-sized alloy on an Al₂O₃ support through a wet-impregnation method, and used it as a high temperature reforming catalyst for the reverse water gas shift reaction (Fig. 11b).²¹⁹ The conversion of CO₂ was greatly increased by fourfold with the modified catalyst.

	Fig. 11 (a) A schematic illustration of the internal reforming layer fed with isooctane and air. Replotted by the authors. (b) Schematic diagram of a SOEC for CO2 reduction. Plotted by the authors. (c) Schematic sketch illustrating exsolved NPs for the CO2 reduction reaction. Plotted by the authors. (d) HRTEM pattern of the Co–Fe alloy NPs and LSCFM substrate. In situ STEM images of LSCFM: (e) after the reduction treatment in H2 at 700 °C, and (f) after the re-oxidation process in O2 at the same temperature. Reproduced with permission from ref. 220. Copyright 2020, John Wiley and Sons. (g) A typical TEM image of PSCFN powders after being reduced. (h) TEM image of the substrate/NP interface. Reproduced with permission from ref. 221. Copyright 2015, Elsevier. (i) Endurance test of 400 h at 800 °C in dry ethanol under 0.8 V after a reduction treatment in H2 for 100 h. Reproduced with permission from ref. 222. Copyright 2018, Elsevier.

Lv et al. gave rise to an atomic-scale insight into the exsolution of Ni-free Co–Fe alloy NPs in La_0.4Sr_0.6Fe_0.7Co_0.2Mo_0.1O_3−δ (LSCFM) for CO₂ electrolysis, as schematically illustrated in Fig. 11c, by employing in situ scanning TEM (STEM, Fig. 11d) and DFT simulations.²²⁰ It was pointed out that the Mo dopants occupying B sites in LSCFM worked as a driving force for the increase of the segregation energies of Co and Fe ions and enhanced the material chemical stability. In situ STEM observations clearly visualized the concurrent exsolution of Co and Fe ions, forming Co–Fe alloy NPs. More interestingly, reversible exsolution and dissolution of Co–Fe NPs were observed, as presented in Fig. 11e and f. The catalytic measurements demonstrated that the construction of a metal–oxide interface displayed high CO₂ adsorption and activation, and should be highly responsible for the enhanced CO₂ electrolysis performance. Yang et al. prepared a Ni-free fuel electrode composed of (Pr_0.4Sr_0.6)₃(Fe_0.85Nb_0.15)₂O₇ (PSFN) by calcining the Pr_0.4Sr_0.6Co_0.2Fe_0.7Nb_0.1O_3−δ composition in H₂ at 900 °C.²²¹ Surprisingly, Co–Fe bi-metallic NPs were obtained, homogenously dispersed on the PSFN surface, as presented in Fig. 11g and h. A La_0.8Sr_0.2Ga_0.83Mg_0.17O_3−δ-supported SOFC with Co–Fe anchored PCFN catalysts generated PPDs of 0.59–0.92 W cm² in humid C₃H₈ at 800–850 °C, together with a stable power generation ability, suggesting a good coking resistance. In order to achieve direct utilization of dry ethanol in SOFCs, Ni–alloy@FeO_x core–shell NPs were fabricated by spray-coating the additional phase onto the Ni–YSZ fuel electrode.²²² No carbon deposition was observed according to their long-term sustainability research results, as shown in Fig. 11i. All the aforementioned literature studies indicated that appropriate design of pristine compositions is important for targeted NPs in order to achieve advanced electrochemical activities.

It is well recognised that the exsolution technology really promotes the development of electrode materials. As reported by Zhang et al., Ni–Fe alloy NPs decorated on Sr₂Fe_1.5Mo_0.5O_6−δ were developed. The catalysts were further modified by fluorine.²²³ Exsolved Ni–Fe NPs greatly changed the morphology of the catalyst surface as well as the valence states that were much helpful for boosting the adsorption behaviours of the electrode. The cell test outcomes supported that both F-substitution and Ni–Fe NPs modification enhanced CO₂ adsorption density by a factor of 2.4 and increased the surface reaction rate constant (k_chem) for CO₂ reduction to a high rank (18.1 × 10⁻⁵ cm s⁻¹), together with an improved oxygen chemical bulk diffusion coefficient (D_chem), which was more than doubled at 800 °C. Meanwhile, R_p decreased by a degree of 52%. At 800 °C, a remarkably high current density of 2.66 A cm⁻² and a sustainability over 140 h were obtained for CO₂ electrolysis in a SOEC under an applied voltage of 1.5 V. Yu and colleagues ehaviored a Fe modified Ni–Fe alloy-based SDC–Ni_1−xFe_xO_y composite.²²⁴ After evaluation in a SOEC, the most suitable Fe ratio was ascertained to be 40%. A SDC–Ni_0.6Fe_0.4 based cell displayed optimal stability and a low polarization impedance. The introduction of Fe was proved to be effective in preventing Ni particles from being coarsened, and hence positive for enhancing the performance of Ni-based fuel electrodes. Liu and co-workers yielded high-performance La_0.6Sr_0.4Fe_0.8Ni_0.2O_3−δ (LSFN) via an exsolution strategy.²²⁵ They also successfully fabricated Fe–Ni bimetallic nanospheres. According to their TEM detection results, the exsolved alloys were found to be regularly anchored on the host oxide. A YSZ-supported reversible electrochemical cell was fabricated and tested, presenting that the electrochemical reaction kinetics for CO₂ electrolysis was greatly promoted with a R_p of 0.272 Ω cm², along with a remarkable current density as high as 1.78 A cm⁻², and a faradaic efficiency value of about 98.8% at 1.6 V and 850 °C. Accordingly, the authors described that the increasing activated reactant and the improved ionic conductivity, together with the elevated electronic conductivity should be accountable for potentiating the CO₂ electrolysis reaction. However, a small amount of SrCO₃ was observed in their work, indicating surface segregation of Sr, which was not further analysed. Then Ding et al. from the same research group proposed a new compound with a composition of (La_0.65Sr_0.3Ce_0.05)_0.9 (Cr_0.5Fe_0.5)_0.85Ni_0.15O_3−δ, and used it for in situ exsolution of Ni–Fe NPs.²²⁶ The optimized composition induced uniformly distributed perovskite-type particles socketed with Ni–Fe NPs. A cell equipped with the tailored composite displayed significantly enhanced electrochemical reaction kinetics, CO production rate, and faradaic efficiency. In addition, using distributions of relaxed time (DRT) calculations, it was depicted that the Ni additive reduced the Fe segregation energy, revealing the co-segregation of Ni and Fe ions.

Since the selectivity of the CO₂ electrolysis reaction reported in the most literature is very low at low temperature,²²⁷ Liu et al. reported a novel strategy for achieving an extremely high selectivity of ∼100% toward CO using a Ni–Fe nanoalloy.²²⁸ In this work, Ni-doped Sr₂Fe_1.4Mo_0.5Ni_0.175O_6−δ (SFM–Ni0.175) was utilized to derive Ni–Fe alloyed NPs with a size of ∼10 nm. Fig. 12a–d show the differences in catalytic mechanisms of SFM-Ni0.175 and a traditional Ni-cermet electrode. In situ diffuse reflectance infrared spectroscopy (DRIFTS) detection showed the suppression of the formation of formate species, which was deemed to be critical for achieving high catalytic selectivity.

	Fig. 12 Comparative diagrams of the promising CO2 reduction pathways over SFM-Ni0.175 (a) and (b) and Ni-cermet-based electrodes (c) and (d). Reproduced with permission from ref. 228. Copyright 2022, Elsevier.

Chen et al. synthesized La_0.5Sr_0.5Co_0.45Fe_0.45Nb_0.1O_3−δ (LSCFN).²²⁹ It was interesting that, once being annealed in 3% H₂O–97% H₂ between 750 and 850 °C, an A₂B₂O₅ brownmillerite structural phase and exsolved Co–Fe alloy NPs were both detected. They drew a conclusion that the best reducing temperature for exsolution reaction was about 800 °C. It was also demonstrated that a prolonged annealing period could make the particle size even larger and increase the amounts of the Fe content and brownmillerite phase, which would result in high ohmic and resistance values. In another work, Chen et al. investigated the phase evolution ehavior of Sr_0.95Ti_0.35Fe_0.6Ni_0.05O_3−δ (STFN) in 3% H₂O–97% H₂ at 750 °C.²³⁰ The in situ exsolved Ni_0.4Fe_0.6 NPs and a new composition with a chemical formula of AB_0.924O_3−δ were observed. The multi-phase fuel electrode demonstrated decreased polarization resistances. Kang et al. used Ni as the substituent of Cu sites in CuFe₂O₄ (CFO). In a reducing atmosphere, the exsolution of nano-size alloy particles was achieved.²³¹ The power generation results demonstrated that a cell equipped with 10 mol% Ni doped CuFe₂O₄ exhibited the highest output (670 mW cm⁻²) on a LSGM electrolyte at 800 °C. The author attributed the enhanced cell performance to the nano-sized Cu–Fe–Ni alloy on the Fe–Cu metal matrix through the exsolution reaction. Based on the same viewpoint, other typical compounds modified by in-detail composition design were used for the surficial exsolution reaction. More representative literature studies are listed in Tables 1–3.

An electrochemical poling strategy was available and effective for the generation of surficial NPs. Chanthanumataporn et al. investigated the exsolution behaviour of NPs from perovskite-type La_0.43Ca_0.37Ni_0.06Ti_0.94O_3−δ (LCTNi) and La_0.43Ca_0.37Ni_0.03Fe_0.03Ti_0.94O_3−δ (LCTNi–Fe) oxides under applied potentials in a CO₂ atmosphere.²³² A strong Ni NP signal was detected showing the exsolution reaction on the LCTNi substrate, while a Ni–Fe alloy was formed on the surface of LCTNi–Fe. Fig. 13a and b provide the EIS curves of the cells tested before and after electrochemical poling treatment. It was indicated that the onset of the exsolution reaction was 2 V. The average particle size of the derived NPs observed in the SEM results was around 30–100 nm, remaining in a large distribution range. The cells with the modified electrodes displayed desirable results with current densities of 0.37 A cm² for LCTNi and 0.48 Acm² for LCTNi–Fe, at 1.5 V in pure CO₂. In SOFC mode fed with dry H₂, PPDs of 0.36 W cm⁻² and 0.43 W cm⁻² for LCTNi and LCTNi–Fe were obtained, respectively. These results demonstrate that electrical reduction is a useful strategy in for facilitating the exsolution reaction of metallic NPs that are helpful for the improvement of CEC performances.

	Fig. 13 Impedance spectra of the cells equipped with LCTNi and LCTNi-Fe electrodes in CO2 as tested at 900 °C: (a) EIS curves of the samples before and after 3 min potential reduction treatment. (b) EIS results at 1.35 V after different potential reduction procedures. Reproduced with permission from ref. 232. Copyright 2019, Elsevier.

In situ observation of the exsolution process in nanoalloy-anchored catalysts is very meaningful for understanding in-depth the intrinsic lattice reconstruction mechanism, but is still scarce till now. Lv et al. contributed more STEM images illustrating the in situ exsolution and dissolution processes of Co–Fe alloy NPs in Co-doped Sr₂Fe_1.5Mo_0.5O_6−δ (SFMC).²³³ The Co–Fe NP-anchored SFMC was demonstrated to be effective for achieving improved CO₂ electrolysis activity and good sustainability.

	Fig. 14 (a) The overall morphology evolution of the fuel electrode during the redox cycling test with/without infiltration on Ni-SCZ (grey: oxide, green: Ni, white: CeO2). Reproduced with permission from ref. 239. Copyright 2016, Elsevier. (b) A schematic of the reduction-induced exsolution reaction in the fuel electrode. Reproduced with permission from ref. 243. Copyright 2020, Elsevier. (c) A diagram illustrating the role of non-stoichiometry in in situ exsolution of NPs in LPBMFO. Reproduced with permission from ref. 244. Copyright 2020, Elsevier.

Actually, metal ion diffusion and phase migration usually occur in the synthesis process of a fuel electrode, which may also be useful for cell optimization.²⁴⁵ Sciazko et al. made an effort to evaluate the morphological evolution behaviour in the Ni–GDC cermet.²⁴⁶ Derived from this report, a clear microstructure reorganization phenomenon was observed for both Ni and GDC particles. Significant migration of GDC NPs was confirmed based on the SEM observation. Moreover, a nano-sized GDC layer was formed on the Ni surface. The migration of the GDC phase was regarded to be of high relevance for the variation of fuel electrode performance. In addition, this work implies that the phase migration on the scale of nano size was indeed deserved to be investigated for the evaluation of SOFC stability.

Reducing the calcination temperature is helpful for scalable production of NPs. Hwang et al. fabricated Ni and Ba(Zr_0.85Y_0.15)O_3−δ (BZY) composite NPs via impregnating a precursor composed of Ni and BZY into the BZY scaffold at the fuel electrode followed by a calcination process at 900 °C.²⁴⁷ The author considered that the low temperature sintering process for the fuel electrode generated the formation of a BZY and Ni nanocomposite with a size of 20–30 nm. A power output of 790 mW cm⁻² at 700 °C was achieved on a 45 µm electrolyte-supported cell. Additionally, the author optimized the fuel electrode composition by infiltrating CeO₂ and Pd catalysts, and finally observed that non-ohmic ASR of the Ni-cermet was improved. In another literature contributed by Li et al., Sr₂Fe_1.5Mo_0.5O_6−δ (SFM) was uniformly deposited onto a porous Y_0.16Zr_0.84O_2−δ scaffold using aqueous metal–ion solutions through infiltration process (in a symmetrical cell).²⁴⁸ The conductivities were investigated at 600–750 °C, apparently displaying an increased tendency as the infiltrated amount of the SFM loading increased. As a result, the ASR values reduced to be 0.153 Ω cm² with an optimized loading of 9.53 wt% in a 1:1 CO–CO₂ mixed gas, and 0.064 Ω cm² at 11.67 wt% in air. In SOFC mode, the MPDs of SFM-loaded cells increased with the loading amount. In electrolysis mode, the cell indicated a current density of 0.618 A cm⁻² under 1.5 V for a 11.67 wt% SFM-loaded cell.

A combination of different nanotechnologies or construction of multi-phase composites can be used for achieving high-performance catalysts. Liu et al. proposed a perovskite composition for solution infiltration into the ceria backbone.²⁴³ In their work, Ba(Ce_0.9Y_0.1)_0.8Ni_0.2O_3−δ (BCYN) was infiltrated onto the Gd_0.1Ce_0.9O_2−δ (GDC) surface. Ni NPs were then exsolved as a decorated catalyst for the fuel gas cracking. As schematically shown in Fig. 14b, a Ni exsolved Ni–BCYN/GDC composite was prepared, showing attractive electrochemical activity and desirable long-term sustainability in the CH₄ fuel. The R_p values of the cell with a Ni–BCYN/GDC fuel electrode were determined to be 0.085 and 0.12 Ω cm² in H₂ and CH₄ at 750 °C, respectively. Song et al. prepared a composite fuel electrode composed of three phases, including Ni NPs, water-storable BaZr_0.4Ce_0.4Y_0.2O_3−δ oxide, and amorphous BaO.²⁴⁹ The multi-phase electrode was obtained by reducing the pristine Ba(Zr_0.4Ce_0.4Y_0.2)_0.8Ni_0.2O_3−δ perovskite in H₂ at 800 °C.

The formation mechanism of the composite NPs still requires to be further investigated. Zhu et al. reported a facile strategy for developing Fe/MnO_x NPs from a (Pr,Ba)₂Mn_2−yFe_yO_5+δ substrate, and provided a promising mechanism for the enhanced CO₂ catalytic activation achieved based on the first-principles calculation method.²⁴⁴ Their results indicated the introduction of A-site deficiencies drove the exsolution of Fe. More interestingly, a transition from a thermodynamically nonspontaneous process to a spontaneous one was observed for the exsolution reaction. The anchored Fe particles were found to be effective for the improvement of the catalytic activities toward CO₂ (638 mA cm⁻² at 1.60 V). The authors attributed the activity promotion to the improvement of chemical adsorption to CO₂ and the enhancement of electron conduction from exsolved NPs. As logically drawn in Fig. 14c, the oxygen and A-site non-stoichiometry were given as x and y-axes in a Cartesian plot. Fe exsolution in layered (Pr,Ba)₂Mn_2−yFe_yO_5+δ (LPBMFO) was highly suppressed since the transformation was driven by the ideal perovskite structure ABO₃ to A₂B₂O₅ stoichiometry. More oxygens should be removed in order to form double-layer LPBMFO once the reduction reaction occurred. As a consequence, for the stoichiometric LPBMFO, MnO_x particles were more easily to be exsolved on the surface after the reduction process, rather than Fe. The situation might be much different for the A-site non-stoichiometry with the A/B ratio below 0.9. Once being treated under a reducing atmosphere, the deficiency of A-site cations (Pr and Ba ions) would work as a driving force to generate B-site exsolution and produce a great number of Fe/MnO composite NPs.

Technical optimization promotes the synthesis of active NPs in CECs. Somacescu et al. synthesized Ni–Ce–Y–Zr-containing oxides by a hydrothermal route, employing hexadecyltrimethylammonium bromide and tripropylamine as templates.²⁵⁰ Ni segregation and the increase of reduced ceria (Ce³⁺) were observed. These features induced more vacancies for better oxygen migration, consequently facilitating carbon removal. Yue et al. carried out the vacuum infiltration of Gd_0.1Ce_0.9O_1.95 (GDC) nitrate solution into a (La_0.75Sr_0.25)_0.97Cr_0.5Mn_0.5O_3±δ (LSCM)/YSZ skeleton, and concurrently introduced a Pd catalyst into the composite.²⁵¹ The nano-structured phase was found to be advantageous to boosting the electrochemical and catalytic properties toward CO₂ reduction thanks to the vacuum infiltration method.

Wet chemical synthesis routes are widely used for the syntheses of NPs because of their advantages in achieving uniform dissolution of metal ions and relatively low phase formation temperature of the aimed materials with nano size. A sol–gel combustion reaction is a typical strategy and has been commonly used toward the syntheses of nano-size powders, as reported in previous literature.^258–261 For the samples obtained via this method, the particle characteristics will be influenced by several factors. Among them, the chelating agent is one crucial factor contributing to the morphology evolution of NPs. Osman et al. experimentally analysed the chelating agent size effect on the crystallization characteristics of BaCe_0.54Zr_0.36Y_0.1O_2−δ, using six different chelating agents.²⁶² The final results indicated the difference in phase composition and morphology characteristics for the specimens obtained by different chelating agents. Wu et al. developed a flame-based gas-phase fabrication method.²⁶³ The combustion process was modified, by using a swirling spray flame burner. The primary particle size was about 20 nm, which was considered to be helpful for achieving an advanced electrolyte. More importantly, this method was proved to be affordable to scale up. Co-precipitation is another technical method that can be used for preparation of NPs. For example, Choolaei et al. presented experimental evidence of using ammonium tartrate as a precipitant for the synthesis of nanocrystalline Gd_0.1Ce_0.9O_2−δ (GDC10) and Gd_0.2Ce_0.8O_2−δ (GDC20).²⁶⁴ NPs with a uniform particle size of 10–30 nm was obtained. Wet chemical synthesis methods also demonstrated strong application potential for constructing functional interlayers in CECs. For instance, Lyu et al. used a hydrothermal approach to surficial modification of doped ceria, which was then employed as an electrolyte barrier layer for a zirconia electrolyte.²⁶⁵ Both R_o and R_p values were improved, evidencing the successful modification of the interlayer by a wet chemical route.

As referred above, a nano-scale composite electrolyte can work as a special type of dual-ion conducting (H⁺/O²⁻) electrolyte yielding superior cell performances. Excitingly, excellent ionic conductivity can even be maintained in the electrolyte at low temperature, for example, below 500 °C.²⁶⁹ Recently, various literature studies were reported to re-address the nature of CECs with nano-scale electrolytes. Doped ceria is a good ionic conductor, while Co/Fe-based perovskites are perfectly mixed conducting oxides. Rauf et al. reported a composite oxide based on Ba_0.5Sr_0.5Co_0.1Fe_0.7Zr_0.1Y_0.1O_3−δ (BSCFZY) and Ca_0.04Ce_0.8Sm_0.16O_2−δ (SCDC).²⁷⁰ They successfully built a heterostructure between them, achieving both an ionic conductivity as high as 0.22 S cm⁻¹ and an excellent PPD of 900 mW cm⁻² at 520 °C. Well construction of hybrid conducting species and a heterointerface in BSCFZY–SCDC was experimentally confirmed. It was pointed out that the creation of a heterointerface in BSCFZY and SCDC could be described by energy band alignment. The author also made verification by various measurements. The functional interface played an important role in supporting a configuration to promote the ionic conduction and to suppress short-circuiting. Consequently, a distinguished configuration greatly reduced the working temperature of a button cell while maintaining a high power output. As confirmed by Yousaf et al., the TPB region was helpful for constructing a mixing transport channel in composite NPs.²⁷¹ In another work, Xia et al. claimed that BaCo_0.4Fe_0.4Zr_0.1Y_0.1O_3−δ (BCFZY) was also an appropriate H⁺–O²⁻–e⁻ triple-conducting electrode for CECs.²⁷² The electrode material was first presented by Duan et al.^14,18 Xia and co-workers investigated its potential for application as an electrolyte by making use of its attractive ionic conductivity while avoiding its electronic conduction through building a BCFZY–ZnO p–n heterostructure, since BCFZY was a typical p-type material while ZnO worked as a n-type semiconductor.²⁷² As shown in Fig. 17a, with this approach, BCFZY could be applied as an appropriate electrolyte, yielding excellent ionic conductivity and high cell output. Additional investigation confirmed the hybrid H⁺/O²⁻ transport abilities of BCFZY–ZnO. The authors then proposed an energy band alignment mechanism to explain the suppression of internal short-circuiting as well as the ionic conductivity improvement of the heterostructure. These results depicted that traditional semiconductors, which were usually used as electrode materials, were also suitable and desirable candidates for electrolytes. In this regard, Xia et al. developed a Li-containing ZnO semiconductor as a SOFC electrolyte.²⁷³ A PPD of 443 mW cm⁻² along with an open circuit voltage (OCV) of 1.07 V were accomplished at 550 °C. Furthermore, an attractive ionic conductivity of 0.05–0.14 S cm⁻¹ was observed at 450–550 °C. The fluorite (Sm-doped CeO₂, SDC) was combined with perovskite (SrTiO₃, STO), and was utilized as the electrolyte, as reported by Cai et al.²⁷⁴ A cell with a 4SDC–6STO (a mass ratio of 4:6) heterostructure exhibited a PPD of 892 mW cm⁻² at 550 °C. Rauf et al. found that, in Ce-doped BaCo_0.2Fe_0.3−xCe_xZr_0.3Y_0.1Tm_0.1O_3−δ (x = 0.1–0.2), the band structure alignment could be facilely regulated by controlling the Ce amount (Fig. 17b).²⁷⁵ A specimen doped with 20% Ce exhibited a dual conduction behaviour, with oxide-ion and protonic conductivities of 0.193 S cm⁻¹ and 0.09 S cm⁻¹ at 530 °C, respectively. At this temperature, a fuel cell with a BaCo_0.2Fe_0.2Ce_0.2Zr_0.3Y_0.1Tm_0.1O_3−δ electrolyte achieved a PPD of 873 mW cm⁻². A similar semiconducting behaviour can be found in Co-doped SrSnO₃.²⁷⁶ An internal electric field was successfully built up, and the ionic transport was greatly improved.

	Fig. 17 (a) Schematic sketch of a p–n heterojunction built at the BCFZY-ZnO interface and the as-proposed mechanism based on the illustration of the energy band alignment. Reproduced with permission from ref. 272. Copyright 2019, Springer Nature. (b) Schematic diagram of the energy band structures of BaCo0.2Fe0.2Ce0.1Tm0.1Zr0.3Y0.1O3−δ and BaCo0.2Fe0.1Ce0.2Tm0.1Zr0.3Y0.1O3−δ. Reproduced with permission from ref. 275. Copyright 2021, Elsevier. (c) A HR-TEM image of BKFC and SNDC. (d) PPD and OCV curves as a function of BKFC content. Reproduced with permission from ref. 277. Copyright 2022, Elsevier.

Many literature supported that the energy band of doped ceria is suitable for construction of a built-in electric field with a great number of semiconductors. Tayyab et al. demonstrated that the energy band alignment between Sm_0.2Ce_0.8O_2−δ (SDC) and LiNi_0.8Co_0.15Al_0.05O₂ matched well, and accordingly combined them together.²⁷⁸ An ionic conductivity as high as 0.12 S cm⁻¹ was achieved at 520 °C, and a remarkable PPD of 0.735 W cm⁻² was achieved at the same temperature. Liu et al. utilized p-type Ba_0.9K_0.1Fe_0.5Co_0.5O_3−δ (BKFC) and n-type Sm_0.075Nd_0.075Ce_0.85O_2−δ (SNDC) and successfully built a p–n junction between them (Fig. 17c).²⁷⁷ An ionic conductivity of 0.14 S cm⁻¹ was achieved and a PPD of 911 mW cm⁻² was recorded at 550 °C (Fig. 17d). Liu et al. built a Schottky barrier in Gd_0.15Ni_0.05Ce_0.8O_2−δ (GNDC) and SnO₂, and yielded a high hybrid conductivity of 0.220 S cm⁻¹ at 550 °C.²⁷⁹ An OCV of 1.026 V and a PPD of 879.4 mW cm⁻² were achieved at this temperature. It is interesting that although nano-sized semiconductor junctions have been widely investigated as electrolytes in SOFCs, very few literature about their application in electrolysis cells was reported. The main reason may be ascribed to the unique porous structure of the semiconductor-containing electrolyte that may not afford high working voltage and efficient gas separation in electrolysis mode.

	Fig. 18 (a) HR-TEM image of an alumina oxide-supported cell structure containing a 70 nm-thick electrolyte. Reproduced with permission from ref. 283. Copyright 2015, American Chemical Society. (b) TEM view of SDC nanocolumns on the STO substrate. (c) Frequency dependence of the real part of ac conductivity as tested in the nano-scale SDC–STO film at different temperatures. (d) Temperature dependence of the conductivity of the SDC–STO film. Reproduced with permission from ref. 285. Copyright 2015, Springer Nature.

In-depth research of the ionic transport behaviours would provide key information on the nature of the electrochemical reactions in ceramic materials. Yang et al. used the first-order reversal curve current–voltage method, which could be applied for local probing of ionic/electronic conduction associated with electrochemical reactions at a 10 nm scale in nano-scaffold SDC–STO membranes.²⁸⁵ It was considered as a practical geometry for energy device miniaturization, which was composed of highly crystalline vertical nanocolumns of doped ceria (Fig. 18b). As a consequence, the ionic conductivity was increased by tenfold higher than that of plain SDC membranes. By employing scanning probe microscopy, it was found that the ion transport channels were both confined to the interface and at SDC nanopillars (Fig. 18c and d). The corresponding results provided more confidence for realizing spatially localized rapid ion transport in oxides.

Till now, it remains a challenging task for fabrication of nano-thin perovskite oxide membranes, and very few research groups concentrated on the construction of ultrathin electrolyte layers with traditional ceramic technologies. The reason may be ascribed to the difficulties in densification of perovskite oxides at high temperature. Most perovskites used in CECs are composed of alkaline earth metal ions. The partial metal ion segregation, the grain crack, etc. make the structure fragile at high temperature. Therefore, there are still significant technical difficulties in efficient construction of a nano-thin perovskite membrane.

Due to partial reduction of Ce⁴⁺ to Ce³⁺, doped ceria works as a mixing conductor (O²⁻ and e⁻) under reducing atmosphere and high temperature conditions, and will induce internal short circuiting when being used as an electrolyte material. To suppress internal electronic leakage, Myung et al. fabricated an ultrathin YSZ membrane as a blocking layer between the fuel electrode and the GDC electrolyte.²⁹¹ They prepared four specimens with different sizes (0–200 nm). The OCV values were boosted to be around 1.1 V, for all the as-modified specimens. The cell with 100 nm-thick YSZ output the largest PPD. Sun et al. ingeniously employed internal Ba diffusion for in situ fabrication of a shell over SDC with a perovskite-type structure.²⁹² TEM observation displayed a core/shell structure. Internal electronic leakage was found to be completely eliminated due to the perovskite-structured shell.

An et al. reported a novel strategy for fabrication of a thin-film three-dimensional (3D) SOFC architecture using the nanosphere lithography (NSL) technique.²⁹³ The ultrathin electrolyte membrane was interposed with a nano-sized catalytic interlayer. A PPD of 1.3 W cm⁻² was achieved at 450 °C in a YSZ-supported SOFC, which was an ultrahigh power output for YSZ-based SOFCs at such an ultralow temperature, even comparable to the most outstanding results up to now. The authors believed that the decrease of ohmic and polarization resistances was caused by the combined effects of usage of an ultrathin electrolyte membrane, increased active areas by the 3D architecture, and enhanced catalytic activities facilitated by the ceria-based interlayers. This work provided valuable reference and practical significance that were very helpful for further design of high-performance SOFCs.

Triple-conducting oxides are famous candidates and emerging air electrodes for low-temperature CECs.⁵⁶ Through nano engineering, a micro-size backbone combined with nano-scale particles could be assembled, such as the infiltration approach, introducing a secondary phase with ionic/electronic conduction, which is helpful for constructing triple-conducting electrodes.²⁹⁵ As displayed in Fig. 19a, Cao et al. prepared a unique SOEC constructed by a nano-layer La_0.6Sr_0.4CoO_3−δ perovskite and a vertically aligned micro-scaled channel air electrode scaffold.⁴⁰ The novel structure strengthened interfacial adherence ability, facilitated oxygen generation stability and accelerated oxygen release. A SOEC with a novel air electrode configuration achieved an extremely high current density of 5.73 A cm⁻², which was a record-breaking performance at that time. Then the novel cell demonstrated durable steam electrolysis under 5.96 A cm⁻² at 1.3 V and 800 °C, corresponding to a H₂ production rate of 2.5 L h⁻¹ cm⁻². Actually, Co-based perovskites are appropriate candidates that are widely used for further treatment via infiltration due to their high catalytic activities. Zheng et al. impregnated La_0.8Sr_0.2Co_0.8Ni_0.2O_3−δ (LSCN) onto the surface of a micro-sized GDC10–(La_0.8Sr_0.2)_0.95MnO_3−δ (LSM) hybrid, working as an air electrode for co-electrolysis of CO₂ and H₂O. The results evidenced that, while being impregnated by LSCN, the cell exhibited elevated performance with a PPD of 1057 mW cm⁻² at 800 °C. While in SOEC mode, the current density reached 1.60 A cm⁻² at 800 °C and 1.5 V, when being fed with H₂O/CO₂ (2:1 in ratio).²⁹⁶ Duan et al. reported Zr–Y co-doped BaCo_0.4Fe_0.4Zr_0.1Y_0.1O_3−δ (BCFZY), and successfully achieved fabrication of a nanostructure using the traditional sintering process.²⁹⁷ They also demonstrated that a nanoparticulated BCFZY catalyst phase was obtained using an infiltration approach.¹⁸ A SOFC yielded a PPD of 455 mW cm⁻² at 500 °C thanks to the boosted electrochemical reaction kinetics facilitated by surficial BCFZY NPs. Since Sm_0.2Ce_0.8O_2−δ (SDC) was an oxygen ion conductor, Saqib et al. impregnated it into the air electrode and prepared a H⁺–O²⁻–e⁻-conducting catalytic electrode.²⁹⁸ A cell with decorated SDC exhibited a reduced R_p from 1.388 to 1.079 Ω cm² in a symmetric configuration at 600 °C. Moreover, an increased current intensity of 1.76 A cm⁻² of a modified cell vs 1.17 A cm⁻² of the pristine cell was achieved at 700 °C and at an applied voltage of 1.5 V fed with the 20% steam–80% air mixture. While working in SOFC mode, the SDC-infiltrated PCFC exhibited an improved MPD due to infiltrated SDC. Saher et al. coated porous nano-particulate Gd₂O₃, CeO₂, and GDC20 onto perovskite-type La_0.58Sr_0.4Co_0.2Fe_0.8O_3−δ (LSCF) ceramics.⁷⁵ As shown in Fig. 19b and c, the oxygen surface exchange kinetics of the specimens were experimentally evaluated at 700–900 °C, followed by the pO₂-step varying in the range of 0.2-0.4 atm. D_chem values remained stable for the coated samples, while the apparent level of k_chem was likely to change with the loading mass and the ionic conductivity of the decorated phase (Fig. 19b). The authors then concluded that partial coverage by Gd₂O₃ or CeO₂ reduced the k_chem values due to their inert non-ionic conductivity inducing surficial blocking effects. Nevertheless, the situation would be much different for GDC20. Uncompleted coverage of LSCF with GDC20 NPs apparently enhanced k_chem up to a factor of 6 relative to the untreated sample. The results from the pulse isotope exchange test supported that the surface exchange reaction on bare LSCF was strongly limited by dissociative adsorption of O₂. An exchange oxygen ion between LSCF and GDC20 was conceived, as revealed in Fig. 19c. In turn, LSCF was able to be infiltrated onto the GDC backbone with a LSCF-containing precursor, forming LSCF NPs with small size.²⁹⁹ In order to improve CO₂ electrolysis performance, Huang et al. infiltrated Ce_0.9M_0.1O_2−δ (M = Fe, Co, Ni) co-catalysts into the La_0.6Sr_0.4Fe_0.5Cr_0.5O_3−δ–Gd_0.2Ce_0.8O_2−δ (LSCrFe–GDC) air electrode.³⁰⁰ The Ce_0.9Co_0.1O_2−δ infiltrated cell outperformed the best performance at 1.5 V and 800 °C (0.652 A cm⁻²). ElS combined with DRT analyses displayed that both adsorption and dissociation processes of CO₂ at TPBs were facilitated by co-catalysts.

	Fig. 19 (a) A sketch of the particular SOEC with a nano/micro-channel air electrode. Reproduced with permission from ref. 40. Copyright 2022, John Wiley and Sons. (b) Arrhenius plots of kchem for pristine LSCF and impregnated specimens. (c) Oxygen exchange mechanism on GDC20 NP-coated LSCF. Reproduced with permission from ref. 75. Copyright 2017, the Royal Society of Chemistry.

Other typical oxides are affordable for construction of active NPs in CECs. For example, a Pr₆O₁₁ infiltrated LaNi_0.6Fe_0.4O₃/Ce_0.9Gd_0.1O₂ (LNF/CGO) composite, demonstrated a low R_p value of 0.074 Ω cm² at 600 °C, as reported by Khoshkalam et al.³⁰¹ The infiltration approach inspired Ge and co-workers. They incorporated Pr₆O₁₁ NPs into (La,Sr)MnO_3−δ (LSM), La_0.6Sr_0.4Co_0.2Fe_0.8O_3−δ (LSCF) and La_0.6Sr_0.4CoO_3−δ (LSC).³⁰² The R_p value of LSM was reduced to be about ten percent of its original value, while for LSCF and LSC, the R_p values were greatly reduced in the range of 600–800 °C as well. Li and colleagues conducted an infiltration treatment of active Y_0.25Bi_0.75O_1.5 into the LSCF electrode.³⁰³ An ultralow temperature of 600 °C was experimentally demonstrated suitable for the generation of a pure material, due to the low phase formation and melting points of bismuth oxides. Moreover, NiO was used in co-catalysts through infiltration. Lee et al. impregnated nickel nitrate solution into the air electrode, and the size of NiO components was observed to be 10–50 nm.³⁰⁴ A low polarization resistance and a high PPD of 780 mW cm⁻² were yielded at 600 °C.

Infiltration of one perovskite into another, forming a composite perovskite, is effective for the electrochemical activity improvement. Wang et al. infiltrated LaCoO₃ (LCO) into the BaZr_0.8Y_0.2O_3−δ (BZY) scaffold.³⁰⁵ Porous BZY backbones were first obtained, and LCO NPs with a size of 50–100 nm were then achieved through the impregnation method. A R_p of 0.56 Ω cm² was obtained at 600 °C, together with a superior long-term sustainability at 600 °C for 900 h. Meanwhile, they infiltrated Ba_0.5Gd_0.8La_0.7Co₂O_3−δ (BGLC) into the BaZr_0.8Y_0.2O_3−δ (BZY20) backbone.³⁰⁶ BGLC was a triple-conducting double perovskite. Then the introduction of BGLC could be helpful for improving the electrochemical reactions of the electrode, including the diffusion of adsorbed oxygen, oxygen reduction, charge transfer, etc. Huang et al. prepared a La_0.6Sr_0.4CoO_3−δ (LSC) decorated La_0.6Sr_0.4Co_0.2Fe_0.8O_3−δ–Ce_0.9Gd_0.1O_2−δ (LSCF–GDC10) composite air electrode for low-temperature SOFCs by the infiltration method.³⁰⁷ The infiltration process led to intrinsic catalytic activity together with high surface area densities (Fig. 20a). As a consequence, the PPDs of the modified cell were doubled, as compared to the pristine ones. Furthermore, active and exceptional stability testing results as long as over 1300 h were obtained at 600 °C, suggesting high catalytic ability and sustainability of the hybrid air electrode. More perovskites are proved to be effective additives. For example, according to Vøllestad’ research on proton ceramic electrolysers, the air electrode layers were infiltrated with a suspension of Ba_0.5Gd_0.8La_0.7Co₂O_6−δ NPs (<100 nm) for achieving advanced electrode activity.³⁰⁸ Ba_0.5Gd_0.8La_0.7Co₂O_6−δ was also able to be employed as an appropriate functional phase while being infiltrated into the BaZr_0.7Ce_0.2Y_0.1O_3−δ scaffold.³⁰⁹ Namgung et al. reported a modification strategy to the infiltration process.³¹⁰ In their work, the introduction of cetrimonium bromide (CTAB) amino acid (glycine) led to infiltration of individual Sm_0.5Sr_0.5CoO_3−δ (SSC) particles on the electrode backbone, forming a 3D network. Consequently, a button cell demonstrated an enhanced property with an output of 1.57 Wcm⁻² at 700 °C, together with a 100 h durability under 1 A cm⁻². In addition, SSC was also employed for infiltration in the La_0.7Sr_0.3FeO_3−δ–BaZr_0.1Ce_0.7Y_0.2O_3−δ (LSF–BZCY) composite to improve the electrochemical performances of the air electrode, as reported by Chen et al.³¹¹ A coating of 15 wt% SSC was found to be suitable for improving the overall cell performance.

	Fig. 20 (a) HR-TEM observation of LSC infiltrated particles. Reproduced with permission from ref. 307. Copyright 2018, Elsevier. (b) TEM image of a dense Pt (7 nm)-coated LSCF backbone. (c) TEM pattern of the CoOx coating, leading to discrete (Co0.98Fe0.02)Ox. (d) TEM observation of discrete CoOx nanograins on the surface of mixed conductor SDC. Reproduced with permission from ref. 312. Copyright 2021, Elsevier. (e) SEM pattern of green fibers before calcination. (f) Typical TEM and STEM (inset) images focusing on LSCF fibers. Reproduced with permission from ref. 313. Copyright 2016, John Wiley and Sons. (g) and (h) SEM images of the electrode mesh with low- and high-magnification. Reproduced with permission from ref. 27. Copyright 2020, Springer Nature.

Co/Fe-based perovskites, such as BaCo_0.4Fe_0.4Zr_0.1Y_0.1O_3−δ (BCFZY), were usually not stable under a harsh humid atmosphere, and showed high reactivity with H₂O and CO₂.³¹⁴ However, the reaction may be partially reversable under particular conditions, and hence this particular behaviour can be used for conversion of derived BCFZY NPs on an electrode surface above 800 °C. In a YSZ-based SOFC, a PPD of 1.61 W cm⁻² was obtained at 700 °C. Meng also observed the in situ regeneration of BCFZY NPs on BCFZY backbones through a reaction with H₂O and CO₂, and characterized it as the air electrode in electrolysis mode.³¹⁵ The current density increased from 0.30 A cm⁻² to about 1.12 A cm⁻² within about 400 h at 600 °C and 1.3 V.

Traditionally, the infiltration procedure toward a nanostructured air electrode is tedious, which often requires several periodic infiltration treatments and high-temperature calcination cycles. Rehman and co-workers provided a facile and scalable urea-based ultrasonic spray infiltration method for fabrication of nano-scale La_0.6Sr_0.4Co_0.2Fe_0.8O_3−δ (LSCF) for SOFCs.³¹⁶ It was experimentally proved that, by using urea as a precipitating agent, a single calcination procedure was enough for the conversion of the infiltrated phase into a desirable catalyst phase. Even a low temperature of calcination (≤900 °C) could produce a durable SOFC. While combining with the ultrasonic spray technique, the urea-assisted impregnation could be easily scaled up for achieving a desirable electrode area. The cell was then analysed by a long-term durability test for 1200 h, and no signs of degradation were found, demonstrating the effectiveness of the advanced technology in production of durable SOFCs. This research also proposed an alternative way toward commercial fabrication of a durable nanostructured air electrode for SOFCs. Chen et al. proposed a feasible strategy to apply a dense coating layer on a mixed electrode, which was utilized for improving the total conductivity and durability of a SOFC air electrode.³¹² As shown in Fig. 20b–d, Pt or CoO_x was used as the co-catalyst for a LSCF/SDC air electrode, realized by the ALD coating strategy. The decorated layers facilitated a strong interaction between the cocatalysts and the backbones. The Pt coating layer was found to be conformal on LSCF grains, while discrete Pt was observed on SDC. For the sample coated by CoO_x, the conformal layer turned to be discrete, both on LSCF and SDC grains. Anyway, ALD coating of cocatalysts reduced the cell ohmic resistance. Tailoring infiltrated solution with coordinating organic agents could further reduce the size of NPs. During metal–ion infiltration, Błaszczak et al. investigated the influence of β-cyclodextrin on metal NPs.³¹⁷ They found that smaller metal oxides could be obtained on the substrate surface, resulting in increased catalytic activity and enhanced durability of the composite catalysts. To date, infiltration has been widely used in the research of air electrodes. Some classical candidates, such as BSCF,³¹⁸etc., can be used for infiltration. For more critical data, a collection of the literature data reported in the last 5 years was compiled and compared with each other, as shown in Tables 1–3.

Apart from infiltration, more efforts were devoted toward effective fabrication of NPs. Pei et al. reported a new self-reconstruction phenomenon in Ba_0.9Co_0.7Fe_0.2Nb_0.1O_3−δ (BCFN), through a water-promoted surface corrosion process.³² The BCFN electrode was naturally converted into a Nb-rich BCFN electrode (µm scale) decorated with Nb-deficient BCFN NPs. In a R-PCEC, the air electrode achieved good performances at 650 °C, such as, a PPD of 1.70 W cm⁻² in fuel cell mode and a current density of 2.8 A cm⁻² at 1.3 V in electrolysis mode. Chen et al. prepared LSCF nanofibers using an electrospinning strategy, as shown in Fig. 20e and f.³¹³ The diameters of nanofibers were about 50 nm with a pore size of around 50 nm, producing high surface areas for the ORR and high ionic conduction enabled by the fiber structure. Then the air electrode equipped with LSCF nanofibers was tested in a SOFC. According to the experimental results, LSCF nanofiber-equipped air electrodes displayed lower R_p and smaller activation energy for the ORR with respect to a commercial LSCF powder or the nano-LSCF powder which was obtained through the crushed fibres by grinding. The unique electrode architecture favourable to efficient mass/charge transport was considered as the key factor for influencing ORR kinetics, as reflected by DRT analyses of the EIS results. Geometric modelling and simulations suggested that longer and thinner fibers facilitated both mass and charge transport. A single cell with LSCF nanofibers exhibited better performance than the cells with the electrode from commercial LSCF powder or the crushed LSCF nanofibers. Moreover, the cells with a nanofiber electrode displayed excellent sustainability at a constant voltage of 0.7 V for about 450 h at 600 °C when being fed with wet H₂.

Another constructive work for advanced PCEC air electrodes was contributed by Ding et al.²⁷ As presented in Fig. 20g and h, a self-architectured mesh-like electrode was synthesized in order to build a highly porous frame for fast mass transport. A nanofiber-structured triple conducting oxide of PrNi_0.5Co_0.5O_3−δ (PNC) was developed as the air electrode, presenting superior electrochemical performance at 400–600 °C. As for the results, the self-sustainable and reversible operation was successfully accomplished through converting the generated hydrogen of the SOEC to electricity by the SOFC without any H₂ supply. The outstanding electrocatalytic activity was attributed to the excellent proton conduction and remarkable hydration behaviour, as claimed by the authors.

In addition, substitution of electrode compositions with anti-sintering elements such as Zr, Mo, W, etc. was effective in reducing the particle size, and can be used to realize the generation of NPs, according to our knowledge. Then NPs might be synthesized at high temperature. In this regard, a traditional sol–gel method is available for nano-sized air electrodes. Lei et al. prepared Sr₂Fe_1.5Mo_0.5O_6−δ (SFM) and evaluated its electrochemical properties as the air electrode for SOECs.³¹⁹ The electrolyte material was BaZr_0.8Y_0.2O_3−δ (BZY), which was much stable in a H₂O-containing atmosphere. SFM powders were synthesized through a glycine and citric acid assisted combustion method. To obtain pure SFM, the as-synthesized ash was heat-treated at 1050 °C for 5 h. BZY was mixed with SFM with a ratio of 1:1 to prepare a composite electrode. SEM observations demonstrated that nano-scale SFM–BZY was successfully yielded. Additionally, high SOEC performance with a current density of 0.21 A cm⁻² and a faradaic efficiency of 63.6% were achieved at 600 °C.

One critical challenge impeding the development of reversible protonic CECs (R-PCECs) is the sluggish ORR and OER kinetics at the air electrode, and compositional/structural reconstruction though nanotechnologies provided alternative pathways toward a superior air electrode. Xu et al. synthesized a PrBaCo_1.6Fe_0.2Nb_0.2O_5+δ air electrode, observed exsolved Nb-deficient PrBaCo_1.6Fe_0.2Nb_0.2−xO_5+δ NPs on the backbone, and made electrochemical assessment both in fuel cell and electrolysis cell modes.³²⁰ The in situ formed NPs and parent perovskite benefited high catalytic activity and durability of the electrode toward efficient ORR and OER at 650 °C, achieving a PPD of 1.059 W cm⁻² (fuel cell), a current density of 2.148 A cm⁻² at 1.3 V (electrolysis cell), and superior cyclability at ±0.5 A cm⁻² in dual mode for 200 h. He et al. reported a one-pot synthesis method in self-constructing NPs for R–PCECs.³²¹ They first used Ba₂Co_1.5Mo_0.25Nb_0.25O_6−δ (BC1.5MN) as the precursor. After being sintered at high temperature, BC1.5MN decomposed into BaCoO_3−δ (BCO, 45 wt%, a single perovskite) and Ba_2−xCo_1.5−xMo_0.5Nb_0.5O_6−δ (BCMN, 55 wt%, a double perovskite). In this composite, only BCO was observed to be crystalized into NPs with a size of 50–100 nm. Its high oxygen ion and electronic conductivities boosted ORR and OER performances. At 650 °C, the cell achieved a PPD of 1.17 W cm⁻² and a current density of 2.04 A cm⁻² at 1.3 V. Zhu et al. reported a new compound of Ba_0.8Gd_0.8Pr_0.4Co₂O_5+δ, which can naturally reconfigure a double-perovskite Ba_0.8Gd_0.8−xPr_0.4Co_2−yO_5+δ (BGPC) and single-perovskite Gd_xCo_yO_3−δ (GCO) NPs on the BGPC surface.³²² As a result, the GCO-decorated BGPC electrode displayed a low R_p of 0.136 Ω cm² at 650 °C in a symmetrical cell. When being employed as an air electrode, the cell displayed a PPD of 0.909 W cm⁻² in a button fuel cell and a current density of 2.336 A cm⁻² at 1.3 V as tested in an electrolysis cell. Niu carried out interesting experimental research on Pr_0.5Ba_0.5CoO_3−δ.³²³ While being applied as the infiltration precursor on LSCF, a thin film with a composition of Pr_1−xBa_xCoO_3−δ and exsolved BaCoO_3−δ NPs were obtained, after sintering the precursor at 800 °C for 2 h in air. The modified surface remarkably enhanced the LSCF air electrode electrochemical properties with an improved surface oxygen exchange rate, superior surface proton diffusion, and rapid H₂O and O₂ dissociation. At 600 °C, R_p of the catalyst coated LSCF air electrode was reduced by a factor of 25 (from 1.09 to 0.043 Ω cm²) in air, together with a lower degradation rate. Consequently, at 600 °C, a single cell with a modified LSCF electrode exhibited a high PPD of 1.04 W cm⁻² in a fuel cell and a remarkable current density of 1.82 A cm⁻² at 1.3 V in electrolysis mode. In addition, Chen et al. employed BaCoO_3−δ (BCO) NPs as the active sites for electrochemical reactions. They coated the Ba(NO₃)₂ solution onto the PrBa_0.8Ca_0.2Co₂O_5+δ (PBCC) backbone, and finally obtained BCO NPs.³²⁴ The PPD of the fuel cell was improved from ∼0.85 to ∼1.15 W cm⁻² at 750 °C, facilitated by the surface modification. Then, Zhou et al. from Prof. Liu’ group reported a PrBa_0.8Ca_0.2Co₂O_5+δ air electrode decorated with BaCoO_3−δ NPs and being annealed in air with oversaturated H₂O (Fig. 21).³²⁵ The cell utilizing the as-designed electrode displayed minimal polarization resistance of about 0.24 Ω cm² at 600 °C, as well as high stability against 3–50% humidified air. A high-performance R–PCEC using PBCC–BCO exhibited a PPD of 1.06 W cm⁻² in a fuel cell and a current density of 1.51 A cm⁻² at 1.3 V in an electrolysis cell at 600 °C. More importantly, the RPCECs demonstrated an extremely strong durability over 1833 h as tested in electrolysis mode.

	Fig. 21 A schematic sketch of an R-PCEC with a BCO-anchored PBCC air electrode operated in different modes. Reproduced with permission from ref. 325. Copyright 2021, American Chemical Society.

	Fig. 28 (a) A TEM image of the LSC–GDC multilayer on the YSZ substrate. (b) A schematic illustration demonstrating the metal ions interdiffusion occurred at LSC–GDC interfaces and the generation of oxygen vacancies. Reproduced with permission from ref. 344. Copyright 2018, Elsevier.

Thermal and chemical mismatch between the electrolyte and the air electrode layers remains to be a challenging issue in CECs, hindering their wide commercialization. A nano engineering approach is considered as an effective way for alleviating the detrimental effect caused by the mismatch. Choi et al. coated a dense PrBa_0.5Sr_0.5Co_1.5Fe_0.5O_5+δ (PBSCF) film (about 100 nm) on the BaZr_0.4Ce_0.4Y_0.1Yb_0.1O_3−δ (BZCYYb4411) electrolyte via PLD.³⁴⁷ Due to the high proton permeability of PBSCF and alleviation of thermal mismatch between different layers, the cell exhibited an ultrahigh PPD (500 mW cm⁻²) at 500 °C. Accordingly, they provided a new concept on realizing thermo-mechanical compatibility between the electrolyte and electrode layers.

Mesoporous materials, a special architecture composed of NPs and nanopores, are characterized by their high specific surface areas, and can be applied to enhance the electrochemical reaction kinetics of an air electrode. Chen carried out an experimental research on a ZrO₂ coated La_1−xSr_xMnO_3−δ/YSZ (LSM/YSZ) backbone employing ALD and thermal treatment.³⁵¹ Nano-size mesopores were formed, as observed in Fig. 29a, making the entire electrode surface open for gas penetration with high ionic conductivity and electrochemical activity. The nano-ionic network was found to be extremely stable after electrochemical operation between 650 and 800 °C for 400 h. A templating approach has been successfully employed in the fabrication of other mesoporous perovskites that can be used as electrodes in CECs.³⁵² In another work, Hernández et al. synthesized a mesoporous GDC20 scaffold (Fig. 29b), and infiltrated La_0.6Sr_0.4Co_0.2Fe_0.8O_3−δ (LSCF) into GDC20 forming a catalyst composite.³⁵³ The fabrication process was as follows: preparing stoichiometric (Gd, Ce-containing nitrates) precursors in ethanol to impregnate KIT–6 and migrating precursors into the pores by capillary condensation and evaporation of the solvent. Then mesoporous GDC20 could be prepared within the template via the decomposition process at 600 °C. Finally, the silica template was removed from the replica by washing the sample with NaOH and water, respectively. It was proved to be an appropriate strategy toward achieving a narrow mesopore size distribution, a nanocrystalline channel, etc., which can be used for improvement of mass transport, electronic conductivity, and charge mobility in the TPB region. A high current density over 1.2 A cm⁻² at 1.4 V was obtained at 750 °C. The total degradation rate of lower than 2% kh⁻¹ at 0.5 A cm⁻² was yielded. The authors made a comparison of the EIS data measured for a test time of 1300 h under OCV conditions. The polarization resistance caused by the electrodes increased in the first 800 h, but then decreased to the initial level before the operation time of 1300 h, which apparently remains unchanged during the whole test period. Anelli et al. developed mesoporous GDC backbones by decorating the functional LSCF phase.³⁵⁴ The modified cells exhibited excellent electrochemical properties achieving an outstanding fuel cell performance and co-electrolysis properties fed with steam and carbon dioxide with a PPD of 1.35 W cm⁻² at 0.7 V and a current density of 1.30 A cm⁻² at 1.3 V, respectively, and at 750 °C. Then, they carried out an experimental investigation on mesostructured GDC with the same method in large-area CECs, using GDC20–LSCF instead of LSCF as the infiltrating composition.³⁵⁵ The sample was tested as the electrode in co-electrolysis mode. The currents as high as 11.2 A at 1.3 V were recorded in an electrolysis cell with 25 cm² area (750 °C), displaying enhanced performances compared to the pristine button cell. The durability results were found to be enhanced either.

	Fig. 29 (a) Cross-sectional TEM micrograph of the porous ZrO2 nanolayer (about 40 nm) on the LSM/YSZ backbone. The inset blue arrows indicate the random distribution of pores. Reproduced with permission from ref. 351. Copyright 2016, Springer Nature. (b) TEM image of the CGO mesoporous powder. Reproduced with permission from ref. 353. Copyright 2018, The Royal Society of Chemistry.

Post-treatment under particular conditions can regenerate nano-size particles/layers/pores, resulting in enhanced cell performances. Su et al. reported an effective top-down strategy for construction of mesoporous BSCF through post treatment of BSCF in H₂O₂.³⁵⁶ Dissolved Ba²⁺ and Sr²⁺ ions were induced by the reaction with H₂O₂, forming mesoporous BSCF. Its specific surface area was increased by about 60 times, and the ORR activity was significantly enhanced, yielding an excellent PPD of about 1800 mW cm⁻² at 800 °C. Yu et al. employed more chemical solutions, including HCl, HNO₃, NaOH, and H₂O₂, and implemented reactivation of the electrodes, resulting in a remarkable improvement in the reduction of ASRs of the electrode.³⁵⁷ Chrzan et al. deposited a mesoporous interlayer between the electrolyte and the electrode by spin coating for 120 seconds at 4000 rpm, and increased the active area for oxygen reduction and charge transfer across the interface.³⁵⁸ All these reports provided good concepts and alternative pathways for decorating a material surface, introducing nano effects, and consequently enhancing the cell electrochemical properties.